parakeet-rs 0.4.3

// Generated by scripts/generate_burn.py from encoder.onnx with burn-onnx 0.22.0. Do not edit.
extern crate alloc;
use burn::prelude::*;
use burn::nn::LayerNorm;
use burn::nn::LayerNormConfig;
use burn::nn::Linear;
use burn::nn::LinearConfig;
use burn::nn::PaddingConfig1d;
use burn::nn::PaddingConfig2d;
use burn::nn::conv::Conv1d;
use burn::nn::conv::Conv1dConfig;
use burn::nn::conv::Conv2d;
use burn::nn::conv::Conv2dConfig;


#[derive(Module, Debug)]
pub struct Submodule1 {
    conv2d1: Conv2d,
    constant423: burn::module::Param<Tensor<1>>,
    constant424: burn::module::Param<Tensor<1>>,
    constant425: burn::module::Param<Tensor<1>>,
    constant696: burn::module::Param<Tensor<2, Int>>,
    conv2d2: Conv2d,
    constant699: burn::module::Param<Tensor<2, Int>>,
    conv2d3: Conv2d,
    conv2d4: Conv2d,
    #[module(skip)]
    device: Device,
}
impl Submodule1 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let conv2d1 = Conv2dConfig::new([1, 256], [3, 3])
            .with_stride([2, 2])
            .with_padding(PaddingConfig2d::Explicit(2, 2, 1, 1))
            .with_dilation([1, 1])
            .with_groups(1)
            .with_bias(true)
            .init(device);
        let constant423: burn::module::Param<Tensor<1>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
            >::from_data(
                burn::tensor::TensorData::from([2f64]),
                (device, burn::tensor::DType::F32),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        let constant424: burn::module::Param<Tensor<1>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
            >::from_data(
                burn::tensor::TensorData::from([1f64]),
                (device, burn::tensor::DType::F32),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        let constant425: burn::module::Param<Tensor<1>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
            >::from_data(
                burn::tensor::TensorData::from([3f64]),
                (device, burn::tensor::DType::F32),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        let constant696: burn::module::Param<Tensor<2, Int>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
                Int,
            >::zeros([1, 33], (device, burn::tensor::DType::I64)),
            device.clone(),
            false,
            [1, 33].into(),
        );
        let conv2d2 = Conv2dConfig::new([256, 256], [3, 3])
            .with_stride([2, 2])
            .with_padding(PaddingConfig2d::Explicit(2, 2, 1, 1))
            .with_dilation([1, 1])
            .with_groups(256)
            .with_bias(true)
            .init(device);
        let constant699: burn::module::Param<Tensor<2, Int>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
                Int,
            >::zeros([1, 17], (device, burn::tensor::DType::I64)),
            device.clone(),
            false,
            [1, 17].into(),
        );
        let conv2d3 = Conv2dConfig::new([256, 256], [1, 1])
            .with_stride([1, 1])
            .with_padding(PaddingConfig2d::Valid)
            .with_dilation([1, 1])
            .with_groups(1)
            .with_bias(true)
            .init(device);
        let conv2d4 = Conv2dConfig::new([256, 256], [3, 3])
            .with_stride([2, 2])
            .with_padding(PaddingConfig2d::Explicit(2, 2, 1, 1))
            .with_dilation([1, 1])
            .with_groups(256)
            .with_bias(true)
            .init(device);
        Self {
            conv2d1,
            constant423,
            constant424,
            constant425,
            constant696,
            conv2d2,
            constant699,
            conv2d3,
            conv2d4,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        processed_signal: Tensor<3>,
        processed_signal_length: Tensor<1, Int>,
    ) -> (Tensor<1, Int>, Tensor<4>) {
        let transpose1_out1 = processed_signal.permute([0, 2, 1]);
        let unsqueeze1_out1: Tensor<4> = transpose1_out1.unsqueeze_dims::<4>(&[1]);
        let cast1_out1 = processed_signal_length.float().cast(burn::tensor::DType::F32);
        let cast2_out1 = cast1_out1.clone().int().cast(burn::tensor::DType::I64);
        let range1_out1 = Tensor::arange(0..65i64, &self.device)
            .cast(burn::tensor::DType::I64)
            .mul_scalar(1i64)
            .add_scalar(0i64);
        let expand1_out1 = {
            let onnx_shape: [i64; 2usize] = [1, 65];
            let input_dims = range1_out1.dims();
            let mut shape = onnx_shape;
            #[allow(clippy::needless_range_loop)]
            for i in 0..1usize {
                let dim_offset = 1usize + i;
                if shape[dim_offset] == 1 && input_dims[i] > 1 {
                    shape[dim_offset] = input_dims[i] as i64;
                }
            }
            range1_out1.expand(shape)
        };
        let unsqueeze2_out1: Tensor<2, Int> = cast2_out1.unsqueeze_dims::<2>(&[1]);
        let less1_out1 = expand1_out1.lower(unsqueeze2_out1);
        let unsqueeze3_out1: Tensor<3, Bool> = less1_out1.unsqueeze_dims::<3>(&[-1]);
        let expand2_out1 = {
            let onnx_shape: [i64; 3usize] = [1, 65, 128];
            let input_dims = unsqueeze3_out1.dims();
            let mut shape = onnx_shape;
            #[allow(clippy::needless_range_loop)]
            for i in 0..3usize {
                let dim_offset = i;
                if shape[dim_offset] == 1 && input_dims[i] > 1 {
                    shape[dim_offset] = input_dims[i] as i64;
                }
            }
            unsqueeze3_out1.expand(shape)
        };
        let cast3_out1 = expand2_out1.float().cast(burn::tensor::DType::F32);
        let unsqueeze4_out1: Tensor<4> = cast3_out1.unsqueeze_dims::<4>(&[1]);
        let mul1_out1 = unsqueeze1_out1.mul(unsqueeze4_out1);
        let conv2d1_out1 = self.conv2d1.forward(mul1_out1);
        let constant423_out1 = self.constant423.val();
        let add1_out1 = cast1_out1.add(constant423_out1.clone());
        let constant424_out1 = self.constant424.val();
        let add2_out1 = add1_out1.add(constant424_out1.clone());
        let constant425_out1 = self.constant425.val();
        let sub1_out1 = add2_out1.sub(constant425_out1.clone());
        let div1_out1 = sub1_out1.div(constant423_out1.clone());
        let floor1_out1 = div1_out1.floor();
        let add3_out1 = floor1_out1.add(constant424_out1.clone());
        let cast4_out1 = add3_out1.clone().int().cast(burn::tensor::DType::I64);
        let unsqueeze5_out1: Tensor<2, Int> = cast4_out1.unsqueeze_dims::<2>(&[1]);
        let constant696_out1 = self.constant696.val();
        let less2_out1 = constant696_out1.lower(unsqueeze5_out1);
        let unsqueeze6_out1: Tensor<3, Bool> = less2_out1.unsqueeze_dims::<3>(&[-1]);
        let expand3_out1 = {
            let onnx_shape: [i64; 3usize] = [1, 33, 65];
            let input_dims = unsqueeze6_out1.dims();
            let mut shape = onnx_shape;
            #[allow(clippy::needless_range_loop)]
            for i in 0..3usize {
                let dim_offset = i;
                if shape[dim_offset] == 1 && input_dims[i] > 1 {
                    shape[dim_offset] = input_dims[i] as i64;
                }
            }
            unsqueeze6_out1.expand(shape)
        };
        let cast5_out1 = expand3_out1.float().cast(burn::tensor::DType::F32);
        let unsqueeze7_out1: Tensor<4> = cast5_out1.unsqueeze_dims::<4>(&[1]);
        let expand4_out1 = {
            let onnx_shape: [i64; 4usize] = [1, 256, 33, 65];
            let input_dims = unsqueeze7_out1.dims();
            let mut shape = onnx_shape;
            #[allow(clippy::needless_range_loop)]
            for i in 0..4usize {
                let dim_offset = i;
                if shape[dim_offset] == 1 && input_dims[i] > 1 {
                    shape[dim_offset] = input_dims[i] as i64;
                }
            }
            unsqueeze7_out1.expand(shape)
        };
        let mul2_out1 = conv2d1_out1.mul(expand4_out1.clone());
        let relu1_out1 = burn::tensor::activation::relu(mul2_out1);
        let mul3_out1 = relu1_out1.mul(expand4_out1);
        let conv2d2_out1 = self.conv2d2.forward(mul3_out1);
        let add4_out1 = add3_out1.add(constant423_out1.clone());
        let add5_out1 = add4_out1.add(constant424_out1.clone());
        let sub2_out1 = add5_out1.sub(constant425_out1.clone());
        let div2_out1 = sub2_out1.div(constant423_out1.clone());
        let floor2_out1 = div2_out1.floor();
        let add6_out1 = floor2_out1.add(constant424_out1.clone());
        let cast6_out1 = add6_out1.clone().int().cast(burn::tensor::DType::I64);
        let unsqueeze8_out1: Tensor<2, Int> = cast6_out1.unsqueeze_dims::<2>(&[1]);
        let constant699_out1 = self.constant699.val();
        let less3_out1 = constant699_out1.lower(unsqueeze8_out1);
        let unsqueeze9_out1: Tensor<3, Bool> = less3_out1.unsqueeze_dims::<3>(&[-1]);
        let expand5_out1 = {
            let onnx_shape: [i64; 3usize] = [1, 17, 33];
            let input_dims = unsqueeze9_out1.dims();
            let mut shape = onnx_shape;
            #[allow(clippy::needless_range_loop)]
            for i in 0..3usize {
                let dim_offset = i;
                if shape[dim_offset] == 1 && input_dims[i] > 1 {
                    shape[dim_offset] = input_dims[i] as i64;
                }
            }
            unsqueeze9_out1.expand(shape)
        };
        let cast7_out1 = expand5_out1.float().cast(burn::tensor::DType::F32);
        let unsqueeze10_out1: Tensor<4> = cast7_out1.unsqueeze_dims::<4>(&[1]);
        let expand6_out1 = {
            let onnx_shape: [i64; 4usize] = [1, 256, 17, 33];
            let input_dims = unsqueeze10_out1.dims();
            let mut shape = onnx_shape;
            #[allow(clippy::needless_range_loop)]
            for i in 0..4usize {
                let dim_offset = i;
                if shape[dim_offset] == 1 && input_dims[i] > 1 {
                    shape[dim_offset] = input_dims[i] as i64;
                }
            }
            unsqueeze10_out1.expand(shape)
        };
        let mul4_out1 = conv2d2_out1.mul(expand6_out1.clone());
        let conv2d3_out1 = self.conv2d3.forward(mul4_out1);
        let mul5_out1 = conv2d3_out1.mul(expand6_out1.clone());
        let relu2_out1 = burn::tensor::activation::relu(mul5_out1);
        let mul6_out1 = relu2_out1.mul(expand6_out1);
        let conv2d4_out1 = self.conv2d4.forward(mul6_out1);
        let add7_out1 = add6_out1.add(constant423_out1.clone());
        let add8_out1 = add7_out1.add(constant424_out1.clone());
        let sub3_out1 = add8_out1.sub(constant425_out1);
        let div3_out1 = sub3_out1.div(constant423_out1);
        let floor3_out1 = div3_out1.floor();
        let add9_out1 = floor3_out1.add(constant424_out1);
        let cast8_out1 = add9_out1.int().cast(burn::tensor::DType::I64);
        (cast8_out1, conv2d4_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule2 {
    constant702: burn::module::Param<Tensor<2, Int>>,
    conv2d5: Conv2d,
    linear1: Linear,
    constant664: burn::module::Param<Tensor<1, Int>>,
    constant659: burn::module::Param<Tensor<1, Int>>,
    constant718: burn::module::Param<Tensor<1, Int>>,
    constant420: burn::module::Param<Tensor<3>>,
    constant430: burn::module::Param<Tensor<1, Bool>>,
    constant433: burn::module::Param<Tensor<1>>,
    #[module(skip)]
    device: Device,
}
impl Submodule2 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let constant702: burn::module::Param<Tensor<2, Int>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
                Int,
            >::zeros([1, 9], (device, burn::tensor::DType::I64)),
            device.clone(),
            false,
            [1, 9].into(),
        );
        let conv2d5 = Conv2dConfig::new([256, 256], [1, 1])
            .with_stride([1, 1])
            .with_padding(PaddingConfig2d::Valid)
            .with_dilation([1, 1])
            .with_groups(1)
            .with_bias(true)
            .init(device);
        let linear1 = LinearConfig::new(4352, 1024).with_bias(true).init(device);
        let constant664: burn::module::Param<Tensor<1, Int>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
                Int,
            >::from_data(
                burn::tensor::TensorData::from([2i64]),
                (device, burn::tensor::DType::I64),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        let constant659: burn::module::Param<Tensor<1, Int>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
                Int,
            >::from_data(
                burn::tensor::TensorData::from([70i64]),
                (device, burn::tensor::DType::I64),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        let constant718: burn::module::Param<Tensor<1, Int>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
                Int,
            >::from_data(
                burn::tensor::TensorData::from([-1i64]),
                (device, burn::tensor::DType::I64),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        let constant420: burn::module::Param<Tensor<3>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                3,
            >::zeros([1, 9999, 1024], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [1, 9999, 1024].into(),
        );
        let constant430: burn::module::Param<Tensor<1, Bool>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
                Bool,
            >::from_data(
                burn::tensor::TensorData::from([true]),
                (device, burn::tensor::DType::Bool(burn::tensor::BoolStore::Native)),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        let constant433: burn::module::Param<Tensor<1>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
            >::from_data(
                burn::tensor::TensorData::from([7f64]),
                (device, burn::tensor::DType::F32),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        Self {
            constant702,
            conv2d5,
            linear1,
            constant664,
            constant659,
            constant718,
            constant420,
            constant430,
            constant433,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        cast8_out1: Tensor<1, Int>,
        conv2d4_out1: Tensor<4>,
        cache_last_channel_len: Tensor<1, Int>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3, Bool>, Tensor<2, Bool>, Tensor<1, Int>) {
        let unsqueeze11_out1: Tensor<2, Int> = cast8_out1
            .clone()
            .unsqueeze_dims::<2>(&[1]);
        let constant702_out1 = self.constant702.val();
        let less4_out1 = constant702_out1.lower(unsqueeze11_out1);
        let unsqueeze12_out1: Tensor<3, Bool> = less4_out1.unsqueeze_dims::<3>(&[-1]);
        let expand7_out1 = {
            let onnx_shape: [i64; 3usize] = [1, 9, 17];
            let input_dims = unsqueeze12_out1.dims();
            let mut shape = onnx_shape;
            #[allow(clippy::needless_range_loop)]
            for i in 0..3usize {
                let dim_offset = i;
                if shape[dim_offset] == 1 && input_dims[i] > 1 {
                    shape[dim_offset] = input_dims[i] as i64;
                }
            }
            unsqueeze12_out1.expand(shape)
        };
        let cast9_out1 = expand7_out1.float().cast(burn::tensor::DType::F32);
        let unsqueeze13_out1: Tensor<4> = cast9_out1.unsqueeze_dims::<4>(&[1]);
        let expand8_out1 = {
            let onnx_shape: [i64; 4usize] = [1, 256, 9, 17];
            let input_dims = unsqueeze13_out1.dims();
            let mut shape = onnx_shape;
            #[allow(clippy::needless_range_loop)]
            for i in 0..4usize {
                let dim_offset = i;
                if shape[dim_offset] == 1 && input_dims[i] > 1 {
                    shape[dim_offset] = input_dims[i] as i64;
                }
            }
            unsqueeze13_out1.expand(shape)
        };
        let mul7_out1 = conv2d4_out1.mul(expand8_out1.clone());
        let conv2d5_out1 = self.conv2d5.forward(mul7_out1);
        let mul8_out1 = conv2d5_out1.mul(expand8_out1.clone());
        let relu3_out1 = burn::tensor::activation::relu(mul8_out1);
        let mul9_out1 = relu3_out1.mul(expand8_out1);
        let transpose2_out1 = mul9_out1.permute([0, 2, 1, 3]);
        let reshape1_out1 = transpose2_out1.reshape([1, 9, 4352]);
        let linear1_out1 = self.linear1.forward(reshape1_out1);
        let slice1_out1 = linear1_out1.slice(s![.., 2.., ..]);
        let constant664_out1 = self.constant664.val();
        let sub4_out1 = cast8_out1.sub(constant664_out1);
        let clip1_out1 = sub4_out1.clamp_min(0f64);
        let constant659_out1 = self.constant659.val();
        let add10_out1 = clip1_out1.clone().add(constant659_out1.clone());
        let constant718_out1 = self.constant718.val();
        let mul10_out1 = cache_last_channel_len.mul(constant718_out1);
        let add11_out1 = mul10_out1.add(constant659_out1);
        let constant420_out1 = self.constant420.val();
        let slice2_out1 = constant420_out1.slice(s![.., 4923..5076, ..]);
        let constant430_out1 = self.constant430.val();
        let expand9_out1 = {
            let shape: [i64; 3] = [1, 77, 77];
            constant430_out1.expand(shape)
        };
        let range2_out1 = Tensor::arange(0..77i64, &self.device)
            .cast(burn::tensor::DType::I32)
            .mul_scalar(1i32)
            .add_scalar(0i32);
        let cast10_out1 = range2_out1.float().cast(burn::tensor::DType::F32);
        let constant433_out1 = self.constant433.val();
        let div4_out1 = cast10_out1.div(constant433_out1);
        let abs1_out1 = div4_out1.clone().abs();
        let floor4_out1 = abs1_out1.floor();
        let sign1_out1 = div4_out1.sign();
        let mul11_out1 = floor4_out1.mul(sign1_out1);
        let cast11_out1 = mul11_out1.int().cast(burn::tensor::DType::I32);
        let unsqueeze14_out1: Tensor<2, Int> = cast11_out1
            .clone()
            .unsqueeze_dims::<2>(&[1]);
        let unsqueeze15_out1: Tensor<2, Int> = cast11_out1.unsqueeze_dims::<2>(&[0]);
        let sub5_out1 = unsqueeze14_out1.sub(unsqueeze15_out1);
        let constant434_out1 = 10i32;
        let lessorequal1_out1 = sub5_out1.clone().lower_equal_elem(constant434_out1);
        let constant431_out1 = 0i32;
        let greaterorequal1_out1 = sub5_out1.greater_equal_elem(constant431_out1);
        let and1_out1 = lessorequal1_out1.bool_and(greaterorequal1_out1);
        let unsqueeze16_out1: Tensor<3, Bool> = and1_out1.unsqueeze_dims::<3>(&[0]);
        let and2_out1 = expand9_out1.bool_and(unsqueeze16_out1);
        let range3_out1 = Tensor::arange(0..77i64, &self.device)
            .cast(burn::tensor::DType::I64)
            .mul_scalar(1i64)
            .add_scalar(0i64);
        let expand10_out1 = {
            let onnx_shape: [i64; 2usize] = [1, 1];
            let input_dims = range3_out1.dims();
            let mut shape = onnx_shape;
            #[allow(clippy::needless_range_loop)]
            for i in 0..1usize {
                let dim_offset = 1usize + i;
                if shape[dim_offset] == 1 && input_dims[i] > 1 {
                    shape[dim_offset] = input_dims[i] as i64;
                }
            }
            range3_out1.expand(shape)
        };
        let unsqueeze17_out1: Tensor<2, Int> = add10_out1.unsqueeze_dims::<2>(&[-1]);
        let less5_out1 = expand10_out1.clone().lower(unsqueeze17_out1);
        let unsqueeze18_out1: Tensor<2, Int> = add11_out1.unsqueeze_dims::<2>(&[-1]);
        let greaterorequal2_out1 = expand10_out1.greater_equal(unsqueeze18_out1);
        let and3_out1 = greaterorequal2_out1.bool_and(less5_out1);
        let unsqueeze19_out1: Tensor<3, Bool> = and3_out1
            .clone()
            .unsqueeze_dims::<3>(&[1]);
        let tile1_out1 = unsqueeze19_out1.repeat(&[1, 77, 1]);
        let transpose3_out1 = tile1_out1.clone().permute([0, 2, 1]);
        let and4_out1 = tile1_out1.bool_and(transpose3_out1);
        let and5_out1 = and4_out1.bool_and(and2_out1);
        let not1_out1 = and5_out1.bool_not();
        let not2_out1 = and3_out1.bool_not();
        let slice3_out1 = not2_out1.slice(s![.., 70..]);
        let slice4_out1 = not1_out1.slice(s![.., 70.., ..]);
        (slice1_out1, slice2_out1, slice4_out1, slice3_out1, clip1_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule3 {
    layernormalization1: LayerNorm,
    linear2: Linear,
    linear3: Linear,
    constant670: burn::module::Param<Tensor<1>>,
    layernormalization2: LayerNorm,
    linear4: Linear,
    linear5: Linear,
    linear6: Linear,
    linear7: Linear,
    constant23: burn::module::Param<Tensor<2>>,
    constant24: burn::module::Param<Tensor<2>>,
    constant672: burn::module::Param<Tensor<1>>,
    linear8: Linear,
    layernormalization3: LayerNorm,
    conv1d1: Conv1d,
    conv1d2: Conv1d,
    layernormalization4: LayerNorm,
    conv1d3: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule3 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization1 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear2 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear3 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let constant670: burn::module::Param<Tensor<1>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
            >::from_data(
                burn::tensor::TensorData::from([0.5f64]),
                (device, burn::tensor::DType::F32),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        let layernormalization2 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear4 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear5 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear6 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear7 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant23: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant24: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant672: burn::module::Param<Tensor<1>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
            >::from_data(
                burn::tensor::TensorData::from([0.0883883461356163f64]),
                (device, burn::tensor::DType::F32),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        let linear8 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization3 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d1 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d2 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization4 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d3 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization1,
            linear2,
            linear3,
            constant670,
            layernormalization2,
            linear4,
            linear5,
            linear6,
            linear7,
            constant23,
            constant24,
            constant672,
            linear8,
            layernormalization3,
            conv1d1,
            conv1d2,
            layernormalization4,
            conv1d3,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice1_out1: Tensor<3>,
        slice2_out1: Tensor<3>,
        slice4_out1: Tensor<3, Bool>,
        slice3_out1: Tensor<2, Bool>,
    ) -> (
        Tensor<3>,
        Tensor<1>,
        Tensor<1>,
        Tensor<4, Bool>,
        f32,
        Tensor<4, Bool>,
        f32,
        Tensor<3, Bool>,
        Tensor<3>,
        Tensor<3>,
    ) {
        let gather1_out1 = {
            let sliced = cache_last_channel.slice(s![0, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather2_out1 = {
            let sliced = cache_last_time.slice(s![0, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization1_out1 = {
            let dtype = slice1_out1.clone().dtype();
            self.layernormalization1
                .forward(slice1_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear2_out1 = self.linear2.forward(layernormalization1_out1);
        let sigmoid1_out1 = burn::tensor::activation::sigmoid(linear2_out1.clone());
        let mul12_out1 = linear2_out1.mul(sigmoid1_out1);
        let linear3_out1 = self.linear3.forward(mul12_out1);
        let constant670_out1 = self.constant670.val();
        let mul13_out1 = linear3_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add12_out1 = slice1_out1.add(mul13_out1);
        let layernormalization2_out1 = {
            let dtype = add12_out1.clone().dtype();
            self.layernormalization2
                .forward(add12_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat1_out1 = burn::tensor::Tensor::cat(
            [gather1_out1.clone(), layernormalization2_out1.clone()].into(),
            1,
        );
        let slice5_out1 = gather1_out1.slice(s![.., 7.., ..]);
        let concat2_out1 = burn::tensor::Tensor::cat(
            [slice5_out1, layernormalization2_out1.clone()].into(),
            1,
        );
        let linear4_out1 = self.linear4.forward(layernormalization2_out1);
        let reshape2_out1 = linear4_out1.reshape([1, -1, 8, 128]);
        let linear5_out1 = self.linear5.forward(concat1_out1.clone());
        let reshape3_out1 = linear5_out1.reshape([1, -1, 8, 128]);
        let linear6_out1 = self.linear6.forward(concat1_out1);
        let reshape4_out1 = linear6_out1.reshape([1, -1, 8, 128]);
        let transpose4_out1 = reshape3_out1.permute([0, 2, 1, 3]);
        let transpose5_out1 = reshape4_out1.permute([0, 2, 1, 3]);
        let linear7_out1 = self.linear7.forward(slice2_out1);
        let reshape5_out1 = linear7_out1.reshape([1, -1, 8, 128]);
        let constant23_out1 = self.constant23.val();
        let add13_out1 = reshape2_out1
            .clone()
            .add((constant23_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose6_out1 = add13_out1.permute([0, 2, 1, 3]);
        let constant24_out1 = self.constant24.val();
        let add14_out1 = reshape2_out1
            .add((constant24_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose7_out1 = add14_out1.permute([0, 2, 1, 3]);
        let transpose8_out1 = reshape5_out1.permute([0, 2, 3, 1]);
        let matmul8_out1 = transpose7_out1.matmul(transpose8_out1);
        let pad1_out1 = matmul8_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape6_out1 = pad1_out1.reshape([1, 8, -1, 7]);
        let slice6_out1 = reshape6_out1.slice(s![.., .., 1.., ..]);
        let reshape7_out1 = slice6_out1.reshape([1, 8, 7, 153]);
        let slice7_out1 = reshape7_out1.slice(s![.., .., .., 0..77]);
        let constant672_out1 = self.constant672.val();
        let mul14_out1 = slice7_out1
            .mul((constant672_out1.clone()).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let unsqueeze20_out1: Tensor<4, Bool> = slice4_out1.unsqueeze_dims::<4>(&[1]);
        let constant444_out1 = -10000f32;
        let where1_out1 = mul14_out1
            .mask_fill(unsqueeze20_out1.clone(), constant444_out1);
        let (attention1_out1,) = {
            let q = transpose6_out1;
            let k = transpose4_out1;
            let v = transpose5_out1;
            let score_bias = where1_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention1_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention1_out1 = {
                let dims = attention1_out1.dims();
                attention1_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention1_out1,)
        };
        let cast12_out1 = unsqueeze20_out1.clone().int().cast(burn::tensor::DType::I64);
        let reducemin1_out1 = {
            cast12_out1.min_dim(3usize).squeeze_dims::<3usize>(&[3])
        };
        let cast13_out1 = reducemin1_out1.bool();
        let unsqueeze21_out1: Tensor<4, Bool> = cast13_out1.unsqueeze_dims::<4>(&[-1]);
        let expand12_out1 = {
            let onnx_shape: [i64; 4usize] = [1, 8, 1, 128];
            let input_dims = unsqueeze21_out1.dims();
            let mut shape = onnx_shape;
            #[allow(clippy::needless_range_loop)]
            for i in 0..4usize {
                let dim_offset = i;
                if shape[dim_offset] == 1 && input_dims[i] > 1 {
                    shape[dim_offset] = input_dims[i] as i64;
                }
            }
            unsqueeze21_out1.expand(shape)
        };
        let constant447_out1 = 0f32;
        let where2_out1 = attention1_out1
            .mask_fill(expand12_out1.clone(), constant447_out1);
        let transpose9_out1 = where2_out1.permute([0, 2, 1, 3]);
        let reshape8_out1 = transpose9_out1.reshape([1, -1, 1024]);
        let linear8_out1 = self.linear8.forward(reshape8_out1);
        let add15_out1 = add12_out1.add(linear8_out1);
        let layernormalization3_out1 = {
            let dtype = add15_out1.clone().dtype();
            self.layernormalization3
                .forward(add15_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose10_out1 = layernormalization3_out1.permute([0, 2, 1]);
        let conv1d1_out1 = self.conv1d1.forward(transpose10_out1);
        let split_tensors = conv1d1_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split1_out1, split1_out2] = split_tensors.try_into().unwrap();
        let sigmoid2_out1 = burn::tensor::activation::sigmoid(split1_out2);
        let mul15_out1 = split1_out1.mul(sigmoid2_out1);
        let unsqueeze22_out1: Tensor<3, Bool> = slice3_out1.unsqueeze_dims::<3>(&[1]);
        let where3_out1 = mul15_out1
            .mask_fill(unsqueeze22_out1.clone(), constant447_out1);
        let concat3_out1 = burn::tensor::Tensor::cat(
            [gather2_out1, where3_out1].into(),
            2,
        );
        let slice8_out1 = concat3_out1.clone().slice(s![.., .., - 8..]);
        let conv1d2_out1 = self.conv1d2.forward(concat3_out1);
        let transpose11_out1 = conv1d2_out1.permute([0, 2, 1]);
        let layernormalization4_out1 = {
            let dtype = transpose11_out1.dtype();
            self.layernormalization4
                .forward(transpose11_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose12_out1 = layernormalization4_out1.permute([0, 2, 1]);
        let sigmoid3_out1 = burn::tensor::activation::sigmoid(transpose12_out1.clone());
        let mul16_out1 = transpose12_out1.mul(sigmoid3_out1);
        let conv1d3_out1 = self.conv1d3.forward(mul16_out1);
        let transpose13_out1 = conv1d3_out1.permute([0, 2, 1]);
        let add16_out1 = add15_out1.add(transpose13_out1);
        (
            add16_out1,
            constant670_out1,
            constant672_out1,
            unsqueeze20_out1,
            constant444_out1,
            expand12_out1,
            constant447_out1,
            unsqueeze22_out1,
            concat2_out1,
            slice8_out1,
        )
    }
}
#[derive(Module, Debug)]
pub struct Submodule4 {
    layernormalization5: LayerNorm,
    linear9: Linear,
    linear10: Linear,
    layernormalization6: LayerNorm,
    layernormalization7: LayerNorm,
    linear11: Linear,
    linear12: Linear,
    layernormalization8: LayerNorm,
    linear13: Linear,
    linear14: Linear,
    linear15: Linear,
    linear16: Linear,
    constant40: burn::module::Param<Tensor<2>>,
    constant41: burn::module::Param<Tensor<2>>,
    linear17: Linear,
    layernormalization9: LayerNorm,
    conv1d4: Conv1d,
    conv1d5: Conv1d,
    layernormalization10: LayerNorm,
    conv1d6: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule4 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization5 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear9 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear10 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization6 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization7 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear11 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear12 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization8 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear13 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear14 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear15 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear16 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant40: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant41: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear17 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization9 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d4 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d5 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization10 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d6 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization5,
            linear9,
            linear10,
            layernormalization6,
            layernormalization7,
            linear11,
            linear12,
            layernormalization8,
            linear13,
            linear14,
            linear15,
            linear16,
            constant40,
            constant41,
            linear17,
            layernormalization9,
            conv1d4,
            conv1d5,
            layernormalization10,
            conv1d6,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add16_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization5_out1 = {
            let dtype = add16_out1.clone().dtype();
            self.layernormalization5
                .forward(add16_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear9_out1 = self.linear9.forward(layernormalization5_out1);
        let sigmoid4_out1 = burn::tensor::activation::sigmoid(linear9_out1.clone());
        let mul17_out1 = linear9_out1.mul(sigmoid4_out1);
        let linear10_out1 = self.linear10.forward(mul17_out1);
        let mul18_out1 = linear10_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add17_out1 = add16_out1.add(mul18_out1);
        let layernormalization6_out1 = {
            let dtype = add17_out1.dtype();
            self.layernormalization6
                .forward(add17_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather3_out1 = {
            let sliced = cache_last_channel.slice(s![1, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather4_out1 = {
            let sliced = cache_last_time.slice(s![1, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization7_out1 = {
            let dtype = layernormalization6_out1.clone().dtype();
            self.layernormalization7
                .forward(layernormalization6_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear11_out1 = self.linear11.forward(layernormalization7_out1);
        let sigmoid5_out1 = burn::tensor::activation::sigmoid(linear11_out1.clone());
        let mul19_out1 = linear11_out1.mul(sigmoid5_out1);
        let linear12_out1 = self.linear12.forward(mul19_out1);
        let mul20_out1 = linear12_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add18_out1 = layernormalization6_out1.add(mul20_out1);
        let layernormalization8_out1 = {
            let dtype = add18_out1.clone().dtype();
            self.layernormalization8
                .forward(add18_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat4_out1 = burn::tensor::Tensor::cat(
            [gather3_out1.clone(), layernormalization8_out1.clone()].into(),
            1,
        );
        let slice9_out1 = gather3_out1.slice(s![.., 7.., ..]);
        let concat5_out1 = burn::tensor::Tensor::cat(
            [slice9_out1, layernormalization8_out1.clone()].into(),
            1,
        );
        let linear13_out1 = self.linear13.forward(layernormalization8_out1);
        let reshape9_out1 = linear13_out1.reshape([1, -1, 8, 128]);
        let linear14_out1 = self.linear14.forward(concat4_out1.clone());
        let reshape10_out1 = linear14_out1.reshape([1, -1, 8, 128]);
        let linear15_out1 = self.linear15.forward(concat4_out1);
        let reshape11_out1 = linear15_out1.reshape([1, -1, 8, 128]);
        let transpose14_out1 = reshape10_out1.permute([0, 2, 1, 3]);
        let transpose15_out1 = reshape11_out1.permute([0, 2, 1, 3]);
        let linear16_out1 = self.linear16.forward(slice2_out1);
        let reshape12_out1 = linear16_out1.reshape([1, -1, 8, 128]);
        let constant40_out1 = self.constant40.val();
        let add19_out1 = reshape9_out1
            .clone()
            .add((constant40_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose16_out1 = add19_out1.permute([0, 2, 1, 3]);
        let constant41_out1 = self.constant41.val();
        let add20_out1 = reshape9_out1
            .add((constant41_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose17_out1 = add20_out1.permute([0, 2, 1, 3]);
        let transpose18_out1 = reshape12_out1.permute([0, 2, 3, 1]);
        let matmul18_out1 = transpose17_out1.matmul(transpose18_out1);
        let pad3_out1 = matmul18_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape13_out1 = pad3_out1.reshape([1, 8, -1, 7]);
        let slice10_out1 = reshape13_out1.slice(s![.., .., 1.., ..]);
        let reshape14_out1 = slice10_out1.reshape([1, 8, 7, 153]);
        let slice11_out1 = reshape14_out1.slice(s![.., .., .., 0..77]);
        let mul21_out1 = slice11_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where4_out1 = mul21_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention2_out1,) = {
            let q = transpose16_out1;
            let k = transpose14_out1;
            let v = transpose15_out1;
            let score_bias = where4_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention2_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention2_out1 = {
                let dims = attention2_out1.dims();
                attention2_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention2_out1,)
        };
        let where5_out1 = attention2_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose19_out1 = where5_out1.permute([0, 2, 1, 3]);
        let reshape15_out1 = transpose19_out1.reshape([1, -1, 1024]);
        let linear17_out1 = self.linear17.forward(reshape15_out1);
        let add21_out1 = add18_out1.add(linear17_out1);
        let layernormalization9_out1 = {
            let dtype = add21_out1.clone().dtype();
            self.layernormalization9
                .forward(add21_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose20_out1 = layernormalization9_out1.permute([0, 2, 1]);
        let conv1d4_out1 = self.conv1d4.forward(transpose20_out1);
        let split_tensors = conv1d4_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split2_out1, split2_out2] = split_tensors.try_into().unwrap();
        let sigmoid6_out1 = burn::tensor::activation::sigmoid(split2_out2);
        let mul22_out1 = split2_out1.mul(sigmoid6_out1);
        let where6_out1 = mul22_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat6_out1 = burn::tensor::Tensor::cat(
            [gather4_out1, where6_out1].into(),
            2,
        );
        let slice12_out1 = concat6_out1.clone().slice(s![.., .., - 8..]);
        let conv1d5_out1 = self.conv1d5.forward(concat6_out1);
        let transpose21_out1 = conv1d5_out1.permute([0, 2, 1]);
        let layernormalization10_out1 = {
            let dtype = transpose21_out1.dtype();
            self.layernormalization10
                .forward(transpose21_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose22_out1 = layernormalization10_out1.permute([0, 2, 1]);
        let sigmoid7_out1 = burn::tensor::activation::sigmoid(transpose22_out1.clone());
        let mul23_out1 = transpose22_out1.mul(sigmoid7_out1);
        let conv1d6_out1 = self.conv1d6.forward(mul23_out1);
        let transpose23_out1 = conv1d6_out1.permute([0, 2, 1]);
        let add22_out1 = add21_out1.add(transpose23_out1);
        (add22_out1, concat5_out1, slice12_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule5 {
    layernormalization11: LayerNorm,
    linear18: Linear,
    linear19: Linear,
    layernormalization12: LayerNorm,
    layernormalization13: LayerNorm,
    linear20: Linear,
    linear21: Linear,
    layernormalization14: LayerNorm,
    linear22: Linear,
    linear23: Linear,
    linear24: Linear,
    linear25: Linear,
    constant57: burn::module::Param<Tensor<2>>,
    constant58: burn::module::Param<Tensor<2>>,
    linear26: Linear,
    layernormalization15: LayerNorm,
    conv1d7: Conv1d,
    conv1d8: Conv1d,
    layernormalization16: LayerNorm,
    conv1d9: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule5 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization11 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear18 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear19 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization12 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization13 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear20 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear21 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization14 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear22 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear23 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear24 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear25 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant57: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant58: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear26 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization15 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d7 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d8 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization16 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d9 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization11,
            linear18,
            linear19,
            layernormalization12,
            layernormalization13,
            linear20,
            linear21,
            layernormalization14,
            linear22,
            linear23,
            linear24,
            linear25,
            constant57,
            constant58,
            linear26,
            layernormalization15,
            conv1d7,
            conv1d8,
            layernormalization16,
            conv1d9,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add22_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization11_out1 = {
            let dtype = add22_out1.clone().dtype();
            self.layernormalization11
                .forward(add22_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear18_out1 = self.linear18.forward(layernormalization11_out1);
        let sigmoid8_out1 = burn::tensor::activation::sigmoid(linear18_out1.clone());
        let mul24_out1 = linear18_out1.mul(sigmoid8_out1);
        let linear19_out1 = self.linear19.forward(mul24_out1);
        let mul25_out1 = linear19_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add23_out1 = add22_out1.add(mul25_out1);
        let layernormalization12_out1 = {
            let dtype = add23_out1.dtype();
            self.layernormalization12
                .forward(add23_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather5_out1 = {
            let sliced = cache_last_channel.slice(s![2, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather6_out1 = {
            let sliced = cache_last_time.slice(s![2, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization13_out1 = {
            let dtype = layernormalization12_out1.clone().dtype();
            self.layernormalization13
                .forward(
                    layernormalization12_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear20_out1 = self.linear20.forward(layernormalization13_out1);
        let sigmoid9_out1 = burn::tensor::activation::sigmoid(linear20_out1.clone());
        let mul26_out1 = linear20_out1.mul(sigmoid9_out1);
        let linear21_out1 = self.linear21.forward(mul26_out1);
        let mul27_out1 = linear21_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add24_out1 = layernormalization12_out1.add(mul27_out1);
        let layernormalization14_out1 = {
            let dtype = add24_out1.clone().dtype();
            self.layernormalization14
                .forward(add24_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat7_out1 = burn::tensor::Tensor::cat(
            [gather5_out1.clone(), layernormalization14_out1.clone()].into(),
            1,
        );
        let slice13_out1 = gather5_out1.slice(s![.., 7.., ..]);
        let concat8_out1 = burn::tensor::Tensor::cat(
            [slice13_out1, layernormalization14_out1.clone()].into(),
            1,
        );
        let linear22_out1 = self.linear22.forward(layernormalization14_out1);
        let reshape16_out1 = linear22_out1.reshape([1, -1, 8, 128]);
        let linear23_out1 = self.linear23.forward(concat7_out1.clone());
        let reshape17_out1 = linear23_out1.reshape([1, -1, 8, 128]);
        let linear24_out1 = self.linear24.forward(concat7_out1);
        let reshape18_out1 = linear24_out1.reshape([1, -1, 8, 128]);
        let transpose24_out1 = reshape17_out1.permute([0, 2, 1, 3]);
        let transpose25_out1 = reshape18_out1.permute([0, 2, 1, 3]);
        let linear25_out1 = self.linear25.forward(slice2_out1);
        let reshape19_out1 = linear25_out1.reshape([1, -1, 8, 128]);
        let constant57_out1 = self.constant57.val();
        let add25_out1 = reshape16_out1
            .clone()
            .add((constant57_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose26_out1 = add25_out1.permute([0, 2, 1, 3]);
        let constant58_out1 = self.constant58.val();
        let add26_out1 = reshape16_out1
            .add((constant58_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose27_out1 = add26_out1.permute([0, 2, 1, 3]);
        let transpose28_out1 = reshape19_out1.permute([0, 2, 3, 1]);
        let matmul28_out1 = transpose27_out1.matmul(transpose28_out1);
        let pad5_out1 = matmul28_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape20_out1 = pad5_out1.reshape([1, 8, -1, 7]);
        let slice14_out1 = reshape20_out1.slice(s![.., .., 1.., ..]);
        let reshape21_out1 = slice14_out1.reshape([1, 8, 7, 153]);
        let slice15_out1 = reshape21_out1.slice(s![.., .., .., 0..77]);
        let mul28_out1 = slice15_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where7_out1 = mul28_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention3_out1,) = {
            let q = transpose26_out1;
            let k = transpose24_out1;
            let v = transpose25_out1;
            let score_bias = where7_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention3_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention3_out1 = {
                let dims = attention3_out1.dims();
                attention3_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention3_out1,)
        };
        let where8_out1 = attention3_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose29_out1 = where8_out1.permute([0, 2, 1, 3]);
        let reshape22_out1 = transpose29_out1.reshape([1, -1, 1024]);
        let linear26_out1 = self.linear26.forward(reshape22_out1);
        let add27_out1 = add24_out1.add(linear26_out1);
        let layernormalization15_out1 = {
            let dtype = add27_out1.clone().dtype();
            self.layernormalization15
                .forward(add27_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose30_out1 = layernormalization15_out1.permute([0, 2, 1]);
        let conv1d7_out1 = self.conv1d7.forward(transpose30_out1);
        let split_tensors = conv1d7_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split3_out1, split3_out2] = split_tensors.try_into().unwrap();
        let sigmoid10_out1 = burn::tensor::activation::sigmoid(split3_out2);
        let mul29_out1 = split3_out1.mul(sigmoid10_out1);
        let where9_out1 = mul29_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat9_out1 = burn::tensor::Tensor::cat(
            [gather6_out1, where9_out1].into(),
            2,
        );
        let slice16_out1 = concat9_out1.clone().slice(s![.., .., - 8..]);
        let conv1d8_out1 = self.conv1d8.forward(concat9_out1);
        let transpose31_out1 = conv1d8_out1.permute([0, 2, 1]);
        let layernormalization16_out1 = {
            let dtype = transpose31_out1.dtype();
            self.layernormalization16
                .forward(transpose31_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose32_out1 = layernormalization16_out1.permute([0, 2, 1]);
        let sigmoid11_out1 = burn::tensor::activation::sigmoid(transpose32_out1.clone());
        let mul30_out1 = transpose32_out1.mul(sigmoid11_out1);
        let conv1d9_out1 = self.conv1d9.forward(mul30_out1);
        let transpose33_out1 = conv1d9_out1.permute([0, 2, 1]);
        let add28_out1 = add27_out1.add(transpose33_out1);
        (add28_out1, concat8_out1, slice16_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule6 {
    layernormalization17: LayerNorm,
    linear27: Linear,
    linear28: Linear,
    layernormalization18: LayerNorm,
    layernormalization19: LayerNorm,
    linear29: Linear,
    linear30: Linear,
    layernormalization20: LayerNorm,
    linear31: Linear,
    linear32: Linear,
    linear33: Linear,
    linear34: Linear,
    constant74: burn::module::Param<Tensor<2>>,
    constant75: burn::module::Param<Tensor<2>>,
    linear35: Linear,
    layernormalization21: LayerNorm,
    conv1d10: Conv1d,
    conv1d11: Conv1d,
    layernormalization22: LayerNorm,
    conv1d12: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule6 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization17 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear27 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear28 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization18 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization19 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear29 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear30 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization20 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear31 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear32 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear33 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear34 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant74: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant75: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear35 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization21 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d10 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d11 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization22 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d12 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization17,
            linear27,
            linear28,
            layernormalization18,
            layernormalization19,
            linear29,
            linear30,
            layernormalization20,
            linear31,
            linear32,
            linear33,
            linear34,
            constant74,
            constant75,
            linear35,
            layernormalization21,
            conv1d10,
            conv1d11,
            layernormalization22,
            conv1d12,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add28_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization17_out1 = {
            let dtype = add28_out1.clone().dtype();
            self.layernormalization17
                .forward(add28_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear27_out1 = self.linear27.forward(layernormalization17_out1);
        let sigmoid12_out1 = burn::tensor::activation::sigmoid(linear27_out1.clone());
        let mul31_out1 = linear27_out1.mul(sigmoid12_out1);
        let linear28_out1 = self.linear28.forward(mul31_out1);
        let mul32_out1 = linear28_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add29_out1 = add28_out1.add(mul32_out1);
        let layernormalization18_out1 = {
            let dtype = add29_out1.dtype();
            self.layernormalization18
                .forward(add29_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather7_out1 = {
            let sliced = cache_last_channel.slice(s![3, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather8_out1 = {
            let sliced = cache_last_time.slice(s![3, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization19_out1 = {
            let dtype = layernormalization18_out1.clone().dtype();
            self.layernormalization19
                .forward(
                    layernormalization18_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear29_out1 = self.linear29.forward(layernormalization19_out1);
        let sigmoid13_out1 = burn::tensor::activation::sigmoid(linear29_out1.clone());
        let mul33_out1 = linear29_out1.mul(sigmoid13_out1);
        let linear30_out1 = self.linear30.forward(mul33_out1);
        let mul34_out1 = linear30_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add30_out1 = layernormalization18_out1.add(mul34_out1);
        let layernormalization20_out1 = {
            let dtype = add30_out1.clone().dtype();
            self.layernormalization20
                .forward(add30_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat10_out1 = burn::tensor::Tensor::cat(
            [gather7_out1.clone(), layernormalization20_out1.clone()].into(),
            1,
        );
        let slice17_out1 = gather7_out1.slice(s![.., 7.., ..]);
        let concat11_out1 = burn::tensor::Tensor::cat(
            [slice17_out1, layernormalization20_out1.clone()].into(),
            1,
        );
        let linear31_out1 = self.linear31.forward(layernormalization20_out1);
        let reshape23_out1 = linear31_out1.reshape([1, -1, 8, 128]);
        let linear32_out1 = self.linear32.forward(concat10_out1.clone());
        let reshape24_out1 = linear32_out1.reshape([1, -1, 8, 128]);
        let linear33_out1 = self.linear33.forward(concat10_out1);
        let reshape25_out1 = linear33_out1.reshape([1, -1, 8, 128]);
        let transpose34_out1 = reshape24_out1.permute([0, 2, 1, 3]);
        let transpose35_out1 = reshape25_out1.permute([0, 2, 1, 3]);
        let linear34_out1 = self.linear34.forward(slice2_out1);
        let reshape26_out1 = linear34_out1.reshape([1, -1, 8, 128]);
        let constant74_out1 = self.constant74.val();
        let add31_out1 = reshape23_out1
            .clone()
            .add((constant74_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose36_out1 = add31_out1.permute([0, 2, 1, 3]);
        let constant75_out1 = self.constant75.val();
        let add32_out1 = reshape23_out1
            .add((constant75_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose37_out1 = add32_out1.permute([0, 2, 1, 3]);
        let transpose38_out1 = reshape26_out1.permute([0, 2, 3, 1]);
        let matmul38_out1 = transpose37_out1.matmul(transpose38_out1);
        let pad7_out1 = matmul38_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape27_out1 = pad7_out1.reshape([1, 8, -1, 7]);
        let slice18_out1 = reshape27_out1.slice(s![.., .., 1.., ..]);
        let reshape28_out1 = slice18_out1.reshape([1, 8, 7, 153]);
        let slice19_out1 = reshape28_out1.slice(s![.., .., .., 0..77]);
        let mul35_out1 = slice19_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where10_out1 = mul35_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention4_out1,) = {
            let q = transpose36_out1;
            let k = transpose34_out1;
            let v = transpose35_out1;
            let score_bias = where10_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention4_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention4_out1 = {
                let dims = attention4_out1.dims();
                attention4_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention4_out1,)
        };
        let where11_out1 = attention4_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose39_out1 = where11_out1.permute([0, 2, 1, 3]);
        let reshape29_out1 = transpose39_out1.reshape([1, -1, 1024]);
        let linear35_out1 = self.linear35.forward(reshape29_out1);
        let add33_out1 = add30_out1.add(linear35_out1);
        let layernormalization21_out1 = {
            let dtype = add33_out1.clone().dtype();
            self.layernormalization21
                .forward(add33_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose40_out1 = layernormalization21_out1.permute([0, 2, 1]);
        let conv1d10_out1 = self.conv1d10.forward(transpose40_out1);
        let split_tensors = conv1d10_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split4_out1, split4_out2] = split_tensors.try_into().unwrap();
        let sigmoid14_out1 = burn::tensor::activation::sigmoid(split4_out2);
        let mul36_out1 = split4_out1.mul(sigmoid14_out1);
        let where12_out1 = mul36_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat12_out1 = burn::tensor::Tensor::cat(
            [gather8_out1, where12_out1].into(),
            2,
        );
        let slice20_out1 = concat12_out1.clone().slice(s![.., .., - 8..]);
        let conv1d11_out1 = self.conv1d11.forward(concat12_out1);
        let transpose41_out1 = conv1d11_out1.permute([0, 2, 1]);
        let layernormalization22_out1 = {
            let dtype = transpose41_out1.dtype();
            self.layernormalization22
                .forward(transpose41_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose42_out1 = layernormalization22_out1.permute([0, 2, 1]);
        let sigmoid15_out1 = burn::tensor::activation::sigmoid(transpose42_out1.clone());
        let mul37_out1 = transpose42_out1.mul(sigmoid15_out1);
        let conv1d12_out1 = self.conv1d12.forward(mul37_out1);
        let transpose43_out1 = conv1d12_out1.permute([0, 2, 1]);
        let add34_out1 = add33_out1.add(transpose43_out1);
        (add34_out1, concat11_out1, slice20_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule7 {
    layernormalization23: LayerNorm,
    linear36: Linear,
    linear37: Linear,
    layernormalization24: LayerNorm,
    layernormalization25: LayerNorm,
    linear38: Linear,
    linear39: Linear,
    layernormalization26: LayerNorm,
    linear40: Linear,
    linear41: Linear,
    linear42: Linear,
    linear43: Linear,
    constant91: burn::module::Param<Tensor<2>>,
    constant92: burn::module::Param<Tensor<2>>,
    linear44: Linear,
    layernormalization27: LayerNorm,
    conv1d13: Conv1d,
    conv1d14: Conv1d,
    layernormalization28: LayerNorm,
    conv1d15: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule7 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization23 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear36 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear37 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization24 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization25 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear38 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear39 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization26 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear40 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear41 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear42 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear43 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant91: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant92: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear44 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization27 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d13 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d14 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization28 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d15 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization23,
            linear36,
            linear37,
            layernormalization24,
            layernormalization25,
            linear38,
            linear39,
            layernormalization26,
            linear40,
            linear41,
            linear42,
            linear43,
            constant91,
            constant92,
            linear44,
            layernormalization27,
            conv1d13,
            conv1d14,
            layernormalization28,
            conv1d15,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add34_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization23_out1 = {
            let dtype = add34_out1.clone().dtype();
            self.layernormalization23
                .forward(add34_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear36_out1 = self.linear36.forward(layernormalization23_out1);
        let sigmoid16_out1 = burn::tensor::activation::sigmoid(linear36_out1.clone());
        let mul38_out1 = linear36_out1.mul(sigmoid16_out1);
        let linear37_out1 = self.linear37.forward(mul38_out1);
        let mul39_out1 = linear37_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add35_out1 = add34_out1.add(mul39_out1);
        let layernormalization24_out1 = {
            let dtype = add35_out1.dtype();
            self.layernormalization24
                .forward(add35_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather9_out1 = {
            let sliced = cache_last_channel.slice(s![4, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather10_out1 = {
            let sliced = cache_last_time.slice(s![4, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization25_out1 = {
            let dtype = layernormalization24_out1.clone().dtype();
            self.layernormalization25
                .forward(
                    layernormalization24_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear38_out1 = self.linear38.forward(layernormalization25_out1);
        let sigmoid17_out1 = burn::tensor::activation::sigmoid(linear38_out1.clone());
        let mul40_out1 = linear38_out1.mul(sigmoid17_out1);
        let linear39_out1 = self.linear39.forward(mul40_out1);
        let mul41_out1 = linear39_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add36_out1 = layernormalization24_out1.add(mul41_out1);
        let layernormalization26_out1 = {
            let dtype = add36_out1.clone().dtype();
            self.layernormalization26
                .forward(add36_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat13_out1 = burn::tensor::Tensor::cat(
            [gather9_out1.clone(), layernormalization26_out1.clone()].into(),
            1,
        );
        let slice21_out1 = gather9_out1.slice(s![.., 7.., ..]);
        let concat14_out1 = burn::tensor::Tensor::cat(
            [slice21_out1, layernormalization26_out1.clone()].into(),
            1,
        );
        let linear40_out1 = self.linear40.forward(layernormalization26_out1);
        let reshape30_out1 = linear40_out1.reshape([1, -1, 8, 128]);
        let linear41_out1 = self.linear41.forward(concat13_out1.clone());
        let reshape31_out1 = linear41_out1.reshape([1, -1, 8, 128]);
        let linear42_out1 = self.linear42.forward(concat13_out1);
        let reshape32_out1 = linear42_out1.reshape([1, -1, 8, 128]);
        let transpose44_out1 = reshape31_out1.permute([0, 2, 1, 3]);
        let transpose45_out1 = reshape32_out1.permute([0, 2, 1, 3]);
        let linear43_out1 = self.linear43.forward(slice2_out1);
        let reshape33_out1 = linear43_out1.reshape([1, -1, 8, 128]);
        let constant91_out1 = self.constant91.val();
        let add37_out1 = reshape30_out1
            .clone()
            .add((constant91_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose46_out1 = add37_out1.permute([0, 2, 1, 3]);
        let constant92_out1 = self.constant92.val();
        let add38_out1 = reshape30_out1
            .add((constant92_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose47_out1 = add38_out1.permute([0, 2, 1, 3]);
        let transpose48_out1 = reshape33_out1.permute([0, 2, 3, 1]);
        let matmul48_out1 = transpose47_out1.matmul(transpose48_out1);
        let pad9_out1 = matmul48_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape34_out1 = pad9_out1.reshape([1, 8, -1, 7]);
        let slice22_out1 = reshape34_out1.slice(s![.., .., 1.., ..]);
        let reshape35_out1 = slice22_out1.reshape([1, 8, 7, 153]);
        let slice23_out1 = reshape35_out1.slice(s![.., .., .., 0..77]);
        let mul42_out1 = slice23_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where13_out1 = mul42_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention5_out1,) = {
            let q = transpose46_out1;
            let k = transpose44_out1;
            let v = transpose45_out1;
            let score_bias = where13_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention5_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention5_out1 = {
                let dims = attention5_out1.dims();
                attention5_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention5_out1,)
        };
        let where14_out1 = attention5_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose49_out1 = where14_out1.permute([0, 2, 1, 3]);
        let reshape36_out1 = transpose49_out1.reshape([1, -1, 1024]);
        let linear44_out1 = self.linear44.forward(reshape36_out1);
        let add39_out1 = add36_out1.add(linear44_out1);
        let layernormalization27_out1 = {
            let dtype = add39_out1.clone().dtype();
            self.layernormalization27
                .forward(add39_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose50_out1 = layernormalization27_out1.permute([0, 2, 1]);
        let conv1d13_out1 = self.conv1d13.forward(transpose50_out1);
        let split_tensors = conv1d13_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split5_out1, split5_out2] = split_tensors.try_into().unwrap();
        let sigmoid18_out1 = burn::tensor::activation::sigmoid(split5_out2);
        let mul43_out1 = split5_out1.mul(sigmoid18_out1);
        let where15_out1 = mul43_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat15_out1 = burn::tensor::Tensor::cat(
            [gather10_out1, where15_out1].into(),
            2,
        );
        let slice24_out1 = concat15_out1.clone().slice(s![.., .., - 8..]);
        let conv1d14_out1 = self.conv1d14.forward(concat15_out1);
        let transpose51_out1 = conv1d14_out1.permute([0, 2, 1]);
        let layernormalization28_out1 = {
            let dtype = transpose51_out1.dtype();
            self.layernormalization28
                .forward(transpose51_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose52_out1 = layernormalization28_out1.permute([0, 2, 1]);
        let sigmoid19_out1 = burn::tensor::activation::sigmoid(transpose52_out1.clone());
        let mul44_out1 = transpose52_out1.mul(sigmoid19_out1);
        let conv1d15_out1 = self.conv1d15.forward(mul44_out1);
        let transpose53_out1 = conv1d15_out1.permute([0, 2, 1]);
        let add40_out1 = add39_out1.add(transpose53_out1);
        (add40_out1, concat14_out1, slice24_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule8 {
    layernormalization29: LayerNorm,
    linear45: Linear,
    linear46: Linear,
    layernormalization30: LayerNorm,
    layernormalization31: LayerNorm,
    linear47: Linear,
    linear48: Linear,
    layernormalization32: LayerNorm,
    linear49: Linear,
    linear50: Linear,
    linear51: Linear,
    linear52: Linear,
    constant108: burn::module::Param<Tensor<2>>,
    constant109: burn::module::Param<Tensor<2>>,
    linear53: Linear,
    layernormalization33: LayerNorm,
    conv1d16: Conv1d,
    conv1d17: Conv1d,
    layernormalization34: LayerNorm,
    conv1d18: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule8 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization29 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear45 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear46 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization30 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization31 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear47 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear48 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization32 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear49 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear50 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear51 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear52 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant108: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant109: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear53 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization33 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d16 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d17 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization34 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d18 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization29,
            linear45,
            linear46,
            layernormalization30,
            layernormalization31,
            linear47,
            linear48,
            layernormalization32,
            linear49,
            linear50,
            linear51,
            linear52,
            constant108,
            constant109,
            linear53,
            layernormalization33,
            conv1d16,
            conv1d17,
            layernormalization34,
            conv1d18,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add40_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization29_out1 = {
            let dtype = add40_out1.clone().dtype();
            self.layernormalization29
                .forward(add40_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear45_out1 = self.linear45.forward(layernormalization29_out1);
        let sigmoid20_out1 = burn::tensor::activation::sigmoid(linear45_out1.clone());
        let mul45_out1 = linear45_out1.mul(sigmoid20_out1);
        let linear46_out1 = self.linear46.forward(mul45_out1);
        let mul46_out1 = linear46_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add41_out1 = add40_out1.add(mul46_out1);
        let layernormalization30_out1 = {
            let dtype = add41_out1.dtype();
            self.layernormalization30
                .forward(add41_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather11_out1 = {
            let sliced = cache_last_channel.slice(s![5, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather12_out1 = {
            let sliced = cache_last_time.slice(s![5, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization31_out1 = {
            let dtype = layernormalization30_out1.clone().dtype();
            self.layernormalization31
                .forward(
                    layernormalization30_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear47_out1 = self.linear47.forward(layernormalization31_out1);
        let sigmoid21_out1 = burn::tensor::activation::sigmoid(linear47_out1.clone());
        let mul47_out1 = linear47_out1.mul(sigmoid21_out1);
        let linear48_out1 = self.linear48.forward(mul47_out1);
        let mul48_out1 = linear48_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add42_out1 = layernormalization30_out1.add(mul48_out1);
        let layernormalization32_out1 = {
            let dtype = add42_out1.clone().dtype();
            self.layernormalization32
                .forward(add42_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat16_out1 = burn::tensor::Tensor::cat(
            [gather11_out1.clone(), layernormalization32_out1.clone()].into(),
            1,
        );
        let slice25_out1 = gather11_out1.slice(s![.., 7.., ..]);
        let concat17_out1 = burn::tensor::Tensor::cat(
            [slice25_out1, layernormalization32_out1.clone()].into(),
            1,
        );
        let linear49_out1 = self.linear49.forward(layernormalization32_out1);
        let reshape37_out1 = linear49_out1.reshape([1, -1, 8, 128]);
        let linear50_out1 = self.linear50.forward(concat16_out1.clone());
        let reshape38_out1 = linear50_out1.reshape([1, -1, 8, 128]);
        let linear51_out1 = self.linear51.forward(concat16_out1);
        let reshape39_out1 = linear51_out1.reshape([1, -1, 8, 128]);
        let transpose54_out1 = reshape38_out1.permute([0, 2, 1, 3]);
        let transpose55_out1 = reshape39_out1.permute([0, 2, 1, 3]);
        let linear52_out1 = self.linear52.forward(slice2_out1);
        let reshape40_out1 = linear52_out1.reshape([1, -1, 8, 128]);
        let constant108_out1 = self.constant108.val();
        let add43_out1 = reshape37_out1
            .clone()
            .add((constant108_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose56_out1 = add43_out1.permute([0, 2, 1, 3]);
        let constant109_out1 = self.constant109.val();
        let add44_out1 = reshape37_out1
            .add((constant109_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose57_out1 = add44_out1.permute([0, 2, 1, 3]);
        let transpose58_out1 = reshape40_out1.permute([0, 2, 3, 1]);
        let matmul58_out1 = transpose57_out1.matmul(transpose58_out1);
        let pad11_out1 = matmul58_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape41_out1 = pad11_out1.reshape([1, 8, -1, 7]);
        let slice26_out1 = reshape41_out1.slice(s![.., .., 1.., ..]);
        let reshape42_out1 = slice26_out1.reshape([1, 8, 7, 153]);
        let slice27_out1 = reshape42_out1.slice(s![.., .., .., 0..77]);
        let mul49_out1 = slice27_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where16_out1 = mul49_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention6_out1,) = {
            let q = transpose56_out1;
            let k = transpose54_out1;
            let v = transpose55_out1;
            let score_bias = where16_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention6_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention6_out1 = {
                let dims = attention6_out1.dims();
                attention6_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention6_out1,)
        };
        let where17_out1 = attention6_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose59_out1 = where17_out1.permute([0, 2, 1, 3]);
        let reshape43_out1 = transpose59_out1.reshape([1, -1, 1024]);
        let linear53_out1 = self.linear53.forward(reshape43_out1);
        let add45_out1 = add42_out1.add(linear53_out1);
        let layernormalization33_out1 = {
            let dtype = add45_out1.clone().dtype();
            self.layernormalization33
                .forward(add45_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose60_out1 = layernormalization33_out1.permute([0, 2, 1]);
        let conv1d16_out1 = self.conv1d16.forward(transpose60_out1);
        let split_tensors = conv1d16_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split6_out1, split6_out2] = split_tensors.try_into().unwrap();
        let sigmoid22_out1 = burn::tensor::activation::sigmoid(split6_out2);
        let mul50_out1 = split6_out1.mul(sigmoid22_out1);
        let where18_out1 = mul50_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat18_out1 = burn::tensor::Tensor::cat(
            [gather12_out1, where18_out1].into(),
            2,
        );
        let slice28_out1 = concat18_out1.clone().slice(s![.., .., - 8..]);
        let conv1d17_out1 = self.conv1d17.forward(concat18_out1);
        let transpose61_out1 = conv1d17_out1.permute([0, 2, 1]);
        let layernormalization34_out1 = {
            let dtype = transpose61_out1.dtype();
            self.layernormalization34
                .forward(transpose61_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose62_out1 = layernormalization34_out1.permute([0, 2, 1]);
        let sigmoid23_out1 = burn::tensor::activation::sigmoid(transpose62_out1.clone());
        let mul51_out1 = transpose62_out1.mul(sigmoid23_out1);
        let conv1d18_out1 = self.conv1d18.forward(mul51_out1);
        let transpose63_out1 = conv1d18_out1.permute([0, 2, 1]);
        let add46_out1 = add45_out1.add(transpose63_out1);
        (add46_out1, concat17_out1, slice28_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule9 {
    layernormalization35: LayerNorm,
    linear54: Linear,
    linear55: Linear,
    layernormalization36: LayerNorm,
    layernormalization37: LayerNorm,
    linear56: Linear,
    linear57: Linear,
    layernormalization38: LayerNorm,
    linear58: Linear,
    linear59: Linear,
    linear60: Linear,
    linear61: Linear,
    constant125: burn::module::Param<Tensor<2>>,
    constant126: burn::module::Param<Tensor<2>>,
    linear62: Linear,
    layernormalization39: LayerNorm,
    conv1d19: Conv1d,
    conv1d20: Conv1d,
    layernormalization40: LayerNorm,
    conv1d21: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule9 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization35 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear54 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear55 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization36 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization37 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear56 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear57 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization38 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear58 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear59 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear60 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear61 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant125: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant126: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear62 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization39 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d19 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d20 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization40 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d21 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization35,
            linear54,
            linear55,
            layernormalization36,
            layernormalization37,
            linear56,
            linear57,
            layernormalization38,
            linear58,
            linear59,
            linear60,
            linear61,
            constant125,
            constant126,
            linear62,
            layernormalization39,
            conv1d19,
            conv1d20,
            layernormalization40,
            conv1d21,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add46_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization35_out1 = {
            let dtype = add46_out1.clone().dtype();
            self.layernormalization35
                .forward(add46_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear54_out1 = self.linear54.forward(layernormalization35_out1);
        let sigmoid24_out1 = burn::tensor::activation::sigmoid(linear54_out1.clone());
        let mul52_out1 = linear54_out1.mul(sigmoid24_out1);
        let linear55_out1 = self.linear55.forward(mul52_out1);
        let mul53_out1 = linear55_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add47_out1 = add46_out1.add(mul53_out1);
        let layernormalization36_out1 = {
            let dtype = add47_out1.dtype();
            self.layernormalization36
                .forward(add47_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather13_out1 = {
            let sliced = cache_last_channel.slice(s![6, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather14_out1 = {
            let sliced = cache_last_time.slice(s![6, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization37_out1 = {
            let dtype = layernormalization36_out1.clone().dtype();
            self.layernormalization37
                .forward(
                    layernormalization36_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear56_out1 = self.linear56.forward(layernormalization37_out1);
        let sigmoid25_out1 = burn::tensor::activation::sigmoid(linear56_out1.clone());
        let mul54_out1 = linear56_out1.mul(sigmoid25_out1);
        let linear57_out1 = self.linear57.forward(mul54_out1);
        let mul55_out1 = linear57_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add48_out1 = layernormalization36_out1.add(mul55_out1);
        let layernormalization38_out1 = {
            let dtype = add48_out1.clone().dtype();
            self.layernormalization38
                .forward(add48_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat19_out1 = burn::tensor::Tensor::cat(
            [gather13_out1.clone(), layernormalization38_out1.clone()].into(),
            1,
        );
        let slice29_out1 = gather13_out1.slice(s![.., 7.., ..]);
        let concat20_out1 = burn::tensor::Tensor::cat(
            [slice29_out1, layernormalization38_out1.clone()].into(),
            1,
        );
        let linear58_out1 = self.linear58.forward(layernormalization38_out1);
        let reshape44_out1 = linear58_out1.reshape([1, -1, 8, 128]);
        let linear59_out1 = self.linear59.forward(concat19_out1.clone());
        let reshape45_out1 = linear59_out1.reshape([1, -1, 8, 128]);
        let linear60_out1 = self.linear60.forward(concat19_out1);
        let reshape46_out1 = linear60_out1.reshape([1, -1, 8, 128]);
        let transpose64_out1 = reshape45_out1.permute([0, 2, 1, 3]);
        let transpose65_out1 = reshape46_out1.permute([0, 2, 1, 3]);
        let linear61_out1 = self.linear61.forward(slice2_out1);
        let reshape47_out1 = linear61_out1.reshape([1, -1, 8, 128]);
        let constant125_out1 = self.constant125.val();
        let add49_out1 = reshape44_out1
            .clone()
            .add((constant125_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose66_out1 = add49_out1.permute([0, 2, 1, 3]);
        let constant126_out1 = self.constant126.val();
        let add50_out1 = reshape44_out1
            .add((constant126_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose67_out1 = add50_out1.permute([0, 2, 1, 3]);
        let transpose68_out1 = reshape47_out1.permute([0, 2, 3, 1]);
        let matmul68_out1 = transpose67_out1.matmul(transpose68_out1);
        let pad13_out1 = matmul68_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape48_out1 = pad13_out1.reshape([1, 8, -1, 7]);
        let slice30_out1 = reshape48_out1.slice(s![.., .., 1.., ..]);
        let reshape49_out1 = slice30_out1.reshape([1, 8, 7, 153]);
        let slice31_out1 = reshape49_out1.slice(s![.., .., .., 0..77]);
        let mul56_out1 = slice31_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where19_out1 = mul56_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention7_out1,) = {
            let q = transpose66_out1;
            let k = transpose64_out1;
            let v = transpose65_out1;
            let score_bias = where19_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention7_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention7_out1 = {
                let dims = attention7_out1.dims();
                attention7_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention7_out1,)
        };
        let where20_out1 = attention7_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose69_out1 = where20_out1.permute([0, 2, 1, 3]);
        let reshape50_out1 = transpose69_out1.reshape([1, -1, 1024]);
        let linear62_out1 = self.linear62.forward(reshape50_out1);
        let add51_out1 = add48_out1.add(linear62_out1);
        let layernormalization39_out1 = {
            let dtype = add51_out1.clone().dtype();
            self.layernormalization39
                .forward(add51_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose70_out1 = layernormalization39_out1.permute([0, 2, 1]);
        let conv1d19_out1 = self.conv1d19.forward(transpose70_out1);
        let split_tensors = conv1d19_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split7_out1, split7_out2] = split_tensors.try_into().unwrap();
        let sigmoid26_out1 = burn::tensor::activation::sigmoid(split7_out2);
        let mul57_out1 = split7_out1.mul(sigmoid26_out1);
        let where21_out1 = mul57_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat21_out1 = burn::tensor::Tensor::cat(
            [gather14_out1, where21_out1].into(),
            2,
        );
        let slice32_out1 = concat21_out1.clone().slice(s![.., .., - 8..]);
        let conv1d20_out1 = self.conv1d20.forward(concat21_out1);
        let transpose71_out1 = conv1d20_out1.permute([0, 2, 1]);
        let layernormalization40_out1 = {
            let dtype = transpose71_out1.dtype();
            self.layernormalization40
                .forward(transpose71_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose72_out1 = layernormalization40_out1.permute([0, 2, 1]);
        let sigmoid27_out1 = burn::tensor::activation::sigmoid(transpose72_out1.clone());
        let mul58_out1 = transpose72_out1.mul(sigmoid27_out1);
        let conv1d21_out1 = self.conv1d21.forward(mul58_out1);
        let transpose73_out1 = conv1d21_out1.permute([0, 2, 1]);
        let add52_out1 = add51_out1.add(transpose73_out1);
        (add52_out1, concat20_out1, slice32_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule10 {
    layernormalization41: LayerNorm,
    linear63: Linear,
    linear64: Linear,
    layernormalization42: LayerNorm,
    layernormalization43: LayerNorm,
    linear65: Linear,
    linear66: Linear,
    layernormalization44: LayerNorm,
    linear67: Linear,
    linear68: Linear,
    linear69: Linear,
    linear70: Linear,
    constant142: burn::module::Param<Tensor<2>>,
    constant143: burn::module::Param<Tensor<2>>,
    linear71: Linear,
    layernormalization45: LayerNorm,
    conv1d22: Conv1d,
    conv1d23: Conv1d,
    layernormalization46: LayerNorm,
    conv1d24: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule10 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization41 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear63 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear64 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization42 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization43 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear65 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear66 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization44 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear67 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear68 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear69 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear70 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant142: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant143: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear71 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization45 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d22 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d23 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization46 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d24 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization41,
            linear63,
            linear64,
            layernormalization42,
            layernormalization43,
            linear65,
            linear66,
            layernormalization44,
            linear67,
            linear68,
            linear69,
            linear70,
            constant142,
            constant143,
            linear71,
            layernormalization45,
            conv1d22,
            conv1d23,
            layernormalization46,
            conv1d24,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add52_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization41_out1 = {
            let dtype = add52_out1.clone().dtype();
            self.layernormalization41
                .forward(add52_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear63_out1 = self.linear63.forward(layernormalization41_out1);
        let sigmoid28_out1 = burn::tensor::activation::sigmoid(linear63_out1.clone());
        let mul59_out1 = linear63_out1.mul(sigmoid28_out1);
        let linear64_out1 = self.linear64.forward(mul59_out1);
        let mul60_out1 = linear64_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add53_out1 = add52_out1.add(mul60_out1);
        let layernormalization42_out1 = {
            let dtype = add53_out1.dtype();
            self.layernormalization42
                .forward(add53_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather15_out1 = {
            let sliced = cache_last_channel.slice(s![7, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather16_out1 = {
            let sliced = cache_last_time.slice(s![7, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization43_out1 = {
            let dtype = layernormalization42_out1.clone().dtype();
            self.layernormalization43
                .forward(
                    layernormalization42_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear65_out1 = self.linear65.forward(layernormalization43_out1);
        let sigmoid29_out1 = burn::tensor::activation::sigmoid(linear65_out1.clone());
        let mul61_out1 = linear65_out1.mul(sigmoid29_out1);
        let linear66_out1 = self.linear66.forward(mul61_out1);
        let mul62_out1 = linear66_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add54_out1 = layernormalization42_out1.add(mul62_out1);
        let layernormalization44_out1 = {
            let dtype = add54_out1.clone().dtype();
            self.layernormalization44
                .forward(add54_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat22_out1 = burn::tensor::Tensor::cat(
            [gather15_out1.clone(), layernormalization44_out1.clone()].into(),
            1,
        );
        let slice33_out1 = gather15_out1.slice(s![.., 7.., ..]);
        let concat23_out1 = burn::tensor::Tensor::cat(
            [slice33_out1, layernormalization44_out1.clone()].into(),
            1,
        );
        let linear67_out1 = self.linear67.forward(layernormalization44_out1);
        let reshape51_out1 = linear67_out1.reshape([1, -1, 8, 128]);
        let linear68_out1 = self.linear68.forward(concat22_out1.clone());
        let reshape52_out1 = linear68_out1.reshape([1, -1, 8, 128]);
        let linear69_out1 = self.linear69.forward(concat22_out1);
        let reshape53_out1 = linear69_out1.reshape([1, -1, 8, 128]);
        let transpose74_out1 = reshape52_out1.permute([0, 2, 1, 3]);
        let transpose75_out1 = reshape53_out1.permute([0, 2, 1, 3]);
        let linear70_out1 = self.linear70.forward(slice2_out1);
        let reshape54_out1 = linear70_out1.reshape([1, -1, 8, 128]);
        let constant142_out1 = self.constant142.val();
        let add55_out1 = reshape51_out1
            .clone()
            .add((constant142_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose76_out1 = add55_out1.permute([0, 2, 1, 3]);
        let constant143_out1 = self.constant143.val();
        let add56_out1 = reshape51_out1
            .add((constant143_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose77_out1 = add56_out1.permute([0, 2, 1, 3]);
        let transpose78_out1 = reshape54_out1.permute([0, 2, 3, 1]);
        let matmul78_out1 = transpose77_out1.matmul(transpose78_out1);
        let pad15_out1 = matmul78_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape55_out1 = pad15_out1.reshape([1, 8, -1, 7]);
        let slice34_out1 = reshape55_out1.slice(s![.., .., 1.., ..]);
        let reshape56_out1 = slice34_out1.reshape([1, 8, 7, 153]);
        let slice35_out1 = reshape56_out1.slice(s![.., .., .., 0..77]);
        let mul63_out1 = slice35_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where22_out1 = mul63_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention8_out1,) = {
            let q = transpose76_out1;
            let k = transpose74_out1;
            let v = transpose75_out1;
            let score_bias = where22_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention8_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention8_out1 = {
                let dims = attention8_out1.dims();
                attention8_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention8_out1,)
        };
        let where23_out1 = attention8_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose79_out1 = where23_out1.permute([0, 2, 1, 3]);
        let reshape57_out1 = transpose79_out1.reshape([1, -1, 1024]);
        let linear71_out1 = self.linear71.forward(reshape57_out1);
        let add57_out1 = add54_out1.add(linear71_out1);
        let layernormalization45_out1 = {
            let dtype = add57_out1.clone().dtype();
            self.layernormalization45
                .forward(add57_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose80_out1 = layernormalization45_out1.permute([0, 2, 1]);
        let conv1d22_out1 = self.conv1d22.forward(transpose80_out1);
        let split_tensors = conv1d22_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split8_out1, split8_out2] = split_tensors.try_into().unwrap();
        let sigmoid30_out1 = burn::tensor::activation::sigmoid(split8_out2);
        let mul64_out1 = split8_out1.mul(sigmoid30_out1);
        let where24_out1 = mul64_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat24_out1 = burn::tensor::Tensor::cat(
            [gather16_out1, where24_out1].into(),
            2,
        );
        let slice36_out1 = concat24_out1.clone().slice(s![.., .., - 8..]);
        let conv1d23_out1 = self.conv1d23.forward(concat24_out1);
        let transpose81_out1 = conv1d23_out1.permute([0, 2, 1]);
        let layernormalization46_out1 = {
            let dtype = transpose81_out1.dtype();
            self.layernormalization46
                .forward(transpose81_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose82_out1 = layernormalization46_out1.permute([0, 2, 1]);
        let sigmoid31_out1 = burn::tensor::activation::sigmoid(transpose82_out1.clone());
        let mul65_out1 = transpose82_out1.mul(sigmoid31_out1);
        let conv1d24_out1 = self.conv1d24.forward(mul65_out1);
        let transpose83_out1 = conv1d24_out1.permute([0, 2, 1]);
        let add58_out1 = add57_out1.add(transpose83_out1);
        (add58_out1, concat23_out1, slice36_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule11 {
    layernormalization47: LayerNorm,
    linear72: Linear,
    linear73: Linear,
    layernormalization48: LayerNorm,
    layernormalization49: LayerNorm,
    linear74: Linear,
    linear75: Linear,
    layernormalization50: LayerNorm,
    linear76: Linear,
    linear77: Linear,
    linear78: Linear,
    linear79: Linear,
    constant159: burn::module::Param<Tensor<2>>,
    constant160: burn::module::Param<Tensor<2>>,
    linear80: Linear,
    layernormalization51: LayerNorm,
    conv1d25: Conv1d,
    conv1d26: Conv1d,
    layernormalization52: LayerNorm,
    conv1d27: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule11 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization47 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear72 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear73 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization48 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization49 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear74 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear75 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization50 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear76 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear77 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear78 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear79 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant159: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant160: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear80 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization51 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d25 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d26 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization52 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d27 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization47,
            linear72,
            linear73,
            layernormalization48,
            layernormalization49,
            linear74,
            linear75,
            layernormalization50,
            linear76,
            linear77,
            linear78,
            linear79,
            constant159,
            constant160,
            linear80,
            layernormalization51,
            conv1d25,
            conv1d26,
            layernormalization52,
            conv1d27,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add58_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization47_out1 = {
            let dtype = add58_out1.clone().dtype();
            self.layernormalization47
                .forward(add58_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear72_out1 = self.linear72.forward(layernormalization47_out1);
        let sigmoid32_out1 = burn::tensor::activation::sigmoid(linear72_out1.clone());
        let mul66_out1 = linear72_out1.mul(sigmoid32_out1);
        let linear73_out1 = self.linear73.forward(mul66_out1);
        let mul67_out1 = linear73_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add59_out1 = add58_out1.add(mul67_out1);
        let layernormalization48_out1 = {
            let dtype = add59_out1.dtype();
            self.layernormalization48
                .forward(add59_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather17_out1 = {
            let sliced = cache_last_channel.slice(s![8, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather18_out1 = {
            let sliced = cache_last_time.slice(s![8, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization49_out1 = {
            let dtype = layernormalization48_out1.clone().dtype();
            self.layernormalization49
                .forward(
                    layernormalization48_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear74_out1 = self.linear74.forward(layernormalization49_out1);
        let sigmoid33_out1 = burn::tensor::activation::sigmoid(linear74_out1.clone());
        let mul68_out1 = linear74_out1.mul(sigmoid33_out1);
        let linear75_out1 = self.linear75.forward(mul68_out1);
        let mul69_out1 = linear75_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add60_out1 = layernormalization48_out1.add(mul69_out1);
        let layernormalization50_out1 = {
            let dtype = add60_out1.clone().dtype();
            self.layernormalization50
                .forward(add60_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat25_out1 = burn::tensor::Tensor::cat(
            [gather17_out1.clone(), layernormalization50_out1.clone()].into(),
            1,
        );
        let slice37_out1 = gather17_out1.slice(s![.., 7.., ..]);
        let concat26_out1 = burn::tensor::Tensor::cat(
            [slice37_out1, layernormalization50_out1.clone()].into(),
            1,
        );
        let linear76_out1 = self.linear76.forward(layernormalization50_out1);
        let reshape58_out1 = linear76_out1.reshape([1, -1, 8, 128]);
        let linear77_out1 = self.linear77.forward(concat25_out1.clone());
        let reshape59_out1 = linear77_out1.reshape([1, -1, 8, 128]);
        let linear78_out1 = self.linear78.forward(concat25_out1);
        let reshape60_out1 = linear78_out1.reshape([1, -1, 8, 128]);
        let transpose84_out1 = reshape59_out1.permute([0, 2, 1, 3]);
        let transpose85_out1 = reshape60_out1.permute([0, 2, 1, 3]);
        let linear79_out1 = self.linear79.forward(slice2_out1);
        let reshape61_out1 = linear79_out1.reshape([1, -1, 8, 128]);
        let constant159_out1 = self.constant159.val();
        let add61_out1 = reshape58_out1
            .clone()
            .add((constant159_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose86_out1 = add61_out1.permute([0, 2, 1, 3]);
        let constant160_out1 = self.constant160.val();
        let add62_out1 = reshape58_out1
            .add((constant160_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose87_out1 = add62_out1.permute([0, 2, 1, 3]);
        let transpose88_out1 = reshape61_out1.permute([0, 2, 3, 1]);
        let matmul88_out1 = transpose87_out1.matmul(transpose88_out1);
        let pad17_out1 = matmul88_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape62_out1 = pad17_out1.reshape([1, 8, -1, 7]);
        let slice38_out1 = reshape62_out1.slice(s![.., .., 1.., ..]);
        let reshape63_out1 = slice38_out1.reshape([1, 8, 7, 153]);
        let slice39_out1 = reshape63_out1.slice(s![.., .., .., 0..77]);
        let mul70_out1 = slice39_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where25_out1 = mul70_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention9_out1,) = {
            let q = transpose86_out1;
            let k = transpose84_out1;
            let v = transpose85_out1;
            let score_bias = where25_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention9_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention9_out1 = {
                let dims = attention9_out1.dims();
                attention9_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention9_out1,)
        };
        let where26_out1 = attention9_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose89_out1 = where26_out1.permute([0, 2, 1, 3]);
        let reshape64_out1 = transpose89_out1.reshape([1, -1, 1024]);
        let linear80_out1 = self.linear80.forward(reshape64_out1);
        let add63_out1 = add60_out1.add(linear80_out1);
        let layernormalization51_out1 = {
            let dtype = add63_out1.clone().dtype();
            self.layernormalization51
                .forward(add63_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose90_out1 = layernormalization51_out1.permute([0, 2, 1]);
        let conv1d25_out1 = self.conv1d25.forward(transpose90_out1);
        let split_tensors = conv1d25_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split9_out1, split9_out2] = split_tensors.try_into().unwrap();
        let sigmoid34_out1 = burn::tensor::activation::sigmoid(split9_out2);
        let mul71_out1 = split9_out1.mul(sigmoid34_out1);
        let where27_out1 = mul71_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat27_out1 = burn::tensor::Tensor::cat(
            [gather18_out1, where27_out1].into(),
            2,
        );
        let slice40_out1 = concat27_out1.clone().slice(s![.., .., - 8..]);
        let conv1d26_out1 = self.conv1d26.forward(concat27_out1);
        let transpose91_out1 = conv1d26_out1.permute([0, 2, 1]);
        let layernormalization52_out1 = {
            let dtype = transpose91_out1.dtype();
            self.layernormalization52
                .forward(transpose91_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose92_out1 = layernormalization52_out1.permute([0, 2, 1]);
        let sigmoid35_out1 = burn::tensor::activation::sigmoid(transpose92_out1.clone());
        let mul72_out1 = transpose92_out1.mul(sigmoid35_out1);
        let conv1d27_out1 = self.conv1d27.forward(mul72_out1);
        let transpose93_out1 = conv1d27_out1.permute([0, 2, 1]);
        let add64_out1 = add63_out1.add(transpose93_out1);
        (add64_out1, concat26_out1, slice40_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule12 {
    layernormalization53: LayerNorm,
    linear81: Linear,
    linear82: Linear,
    layernormalization54: LayerNorm,
    layernormalization55: LayerNorm,
    linear83: Linear,
    linear84: Linear,
    layernormalization56: LayerNorm,
    linear85: Linear,
    linear86: Linear,
    linear87: Linear,
    linear88: Linear,
    constant176: burn::module::Param<Tensor<2>>,
    constant177: burn::module::Param<Tensor<2>>,
    linear89: Linear,
    layernormalization57: LayerNorm,
    conv1d28: Conv1d,
    conv1d29: Conv1d,
    layernormalization58: LayerNorm,
    conv1d30: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule12 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization53 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear81 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear82 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization54 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization55 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear83 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear84 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization56 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear85 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear86 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear87 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear88 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant176: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant177: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear89 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization57 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d28 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d29 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization58 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d30 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization53,
            linear81,
            linear82,
            layernormalization54,
            layernormalization55,
            linear83,
            linear84,
            layernormalization56,
            linear85,
            linear86,
            linear87,
            linear88,
            constant176,
            constant177,
            linear89,
            layernormalization57,
            conv1d28,
            conv1d29,
            layernormalization58,
            conv1d30,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add64_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization53_out1 = {
            let dtype = add64_out1.clone().dtype();
            self.layernormalization53
                .forward(add64_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear81_out1 = self.linear81.forward(layernormalization53_out1);
        let sigmoid36_out1 = burn::tensor::activation::sigmoid(linear81_out1.clone());
        let mul73_out1 = linear81_out1.mul(sigmoid36_out1);
        let linear82_out1 = self.linear82.forward(mul73_out1);
        let mul74_out1 = linear82_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add65_out1 = add64_out1.add(mul74_out1);
        let layernormalization54_out1 = {
            let dtype = add65_out1.dtype();
            self.layernormalization54
                .forward(add65_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather19_out1 = {
            let sliced = cache_last_channel.slice(s![9, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather20_out1 = {
            let sliced = cache_last_time.slice(s![9, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization55_out1 = {
            let dtype = layernormalization54_out1.clone().dtype();
            self.layernormalization55
                .forward(
                    layernormalization54_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear83_out1 = self.linear83.forward(layernormalization55_out1);
        let sigmoid37_out1 = burn::tensor::activation::sigmoid(linear83_out1.clone());
        let mul75_out1 = linear83_out1.mul(sigmoid37_out1);
        let linear84_out1 = self.linear84.forward(mul75_out1);
        let mul76_out1 = linear84_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add66_out1 = layernormalization54_out1.add(mul76_out1);
        let layernormalization56_out1 = {
            let dtype = add66_out1.clone().dtype();
            self.layernormalization56
                .forward(add66_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat28_out1 = burn::tensor::Tensor::cat(
            [gather19_out1.clone(), layernormalization56_out1.clone()].into(),
            1,
        );
        let slice41_out1 = gather19_out1.slice(s![.., 7.., ..]);
        let concat29_out1 = burn::tensor::Tensor::cat(
            [slice41_out1, layernormalization56_out1.clone()].into(),
            1,
        );
        let linear85_out1 = self.linear85.forward(layernormalization56_out1);
        let reshape65_out1 = linear85_out1.reshape([1, -1, 8, 128]);
        let linear86_out1 = self.linear86.forward(concat28_out1.clone());
        let reshape66_out1 = linear86_out1.reshape([1, -1, 8, 128]);
        let linear87_out1 = self.linear87.forward(concat28_out1);
        let reshape67_out1 = linear87_out1.reshape([1, -1, 8, 128]);
        let transpose94_out1 = reshape66_out1.permute([0, 2, 1, 3]);
        let transpose95_out1 = reshape67_out1.permute([0, 2, 1, 3]);
        let linear88_out1 = self.linear88.forward(slice2_out1);
        let reshape68_out1 = linear88_out1.reshape([1, -1, 8, 128]);
        let constant176_out1 = self.constant176.val();
        let add67_out1 = reshape65_out1
            .clone()
            .add((constant176_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose96_out1 = add67_out1.permute([0, 2, 1, 3]);
        let constant177_out1 = self.constant177.val();
        let add68_out1 = reshape65_out1
            .add((constant177_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose97_out1 = add68_out1.permute([0, 2, 1, 3]);
        let transpose98_out1 = reshape68_out1.permute([0, 2, 3, 1]);
        let matmul98_out1 = transpose97_out1.matmul(transpose98_out1);
        let pad19_out1 = matmul98_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape69_out1 = pad19_out1.reshape([1, 8, -1, 7]);
        let slice42_out1 = reshape69_out1.slice(s![.., .., 1.., ..]);
        let reshape70_out1 = slice42_out1.reshape([1, 8, 7, 153]);
        let slice43_out1 = reshape70_out1.slice(s![.., .., .., 0..77]);
        let mul77_out1 = slice43_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where28_out1 = mul77_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention10_out1,) = {
            let q = transpose96_out1;
            let k = transpose94_out1;
            let v = transpose95_out1;
            let score_bias = where28_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention10_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention10_out1 = {
                let dims = attention10_out1.dims();
                attention10_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention10_out1,)
        };
        let where29_out1 = attention10_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose99_out1 = where29_out1.permute([0, 2, 1, 3]);
        let reshape71_out1 = transpose99_out1.reshape([1, -1, 1024]);
        let linear89_out1 = self.linear89.forward(reshape71_out1);
        let add69_out1 = add66_out1.add(linear89_out1);
        let layernormalization57_out1 = {
            let dtype = add69_out1.clone().dtype();
            self.layernormalization57
                .forward(add69_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose100_out1 = layernormalization57_out1.permute([0, 2, 1]);
        let conv1d28_out1 = self.conv1d28.forward(transpose100_out1);
        let split_tensors = conv1d28_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split10_out1, split10_out2] = split_tensors.try_into().unwrap();
        let sigmoid38_out1 = burn::tensor::activation::sigmoid(split10_out2);
        let mul78_out1 = split10_out1.mul(sigmoid38_out1);
        let where30_out1 = mul78_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat30_out1 = burn::tensor::Tensor::cat(
            [gather20_out1, where30_out1].into(),
            2,
        );
        let slice44_out1 = concat30_out1.clone().slice(s![.., .., - 8..]);
        let conv1d29_out1 = self.conv1d29.forward(concat30_out1);
        let transpose101_out1 = conv1d29_out1.permute([0, 2, 1]);
        let layernormalization58_out1 = {
            let dtype = transpose101_out1.dtype();
            self.layernormalization58
                .forward(transpose101_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose102_out1 = layernormalization58_out1.permute([0, 2, 1]);
        let sigmoid39_out1 = burn::tensor::activation::sigmoid(
            transpose102_out1.clone(),
        );
        let mul79_out1 = transpose102_out1.mul(sigmoid39_out1);
        let conv1d30_out1 = self.conv1d30.forward(mul79_out1);
        let transpose103_out1 = conv1d30_out1.permute([0, 2, 1]);
        let add70_out1 = add69_out1.add(transpose103_out1);
        (add70_out1, concat29_out1, slice44_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule13 {
    layernormalization59: LayerNorm,
    linear90: Linear,
    linear91: Linear,
    layernormalization60: LayerNorm,
    layernormalization61: LayerNorm,
    linear92: Linear,
    linear93: Linear,
    layernormalization62: LayerNorm,
    linear94: Linear,
    linear95: Linear,
    linear96: Linear,
    linear97: Linear,
    constant193: burn::module::Param<Tensor<2>>,
    constant194: burn::module::Param<Tensor<2>>,
    linear98: Linear,
    layernormalization63: LayerNorm,
    conv1d31: Conv1d,
    conv1d32: Conv1d,
    layernormalization64: LayerNorm,
    conv1d33: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule13 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization59 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear90 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear91 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization60 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization61 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear92 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear93 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization62 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear94 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear95 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear96 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear97 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant193: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant194: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear98 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization63 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d31 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d32 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization64 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d33 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization59,
            linear90,
            linear91,
            layernormalization60,
            layernormalization61,
            linear92,
            linear93,
            layernormalization62,
            linear94,
            linear95,
            linear96,
            linear97,
            constant193,
            constant194,
            linear98,
            layernormalization63,
            conv1d31,
            conv1d32,
            layernormalization64,
            conv1d33,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add70_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization59_out1 = {
            let dtype = add70_out1.clone().dtype();
            self.layernormalization59
                .forward(add70_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear90_out1 = self.linear90.forward(layernormalization59_out1);
        let sigmoid40_out1 = burn::tensor::activation::sigmoid(linear90_out1.clone());
        let mul80_out1 = linear90_out1.mul(sigmoid40_out1);
        let linear91_out1 = self.linear91.forward(mul80_out1);
        let mul81_out1 = linear91_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add71_out1 = add70_out1.add(mul81_out1);
        let layernormalization60_out1 = {
            let dtype = add71_out1.dtype();
            self.layernormalization60
                .forward(add71_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather21_out1 = {
            let sliced = cache_last_channel.slice(s![10, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather22_out1 = {
            let sliced = cache_last_time.slice(s![10, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization61_out1 = {
            let dtype = layernormalization60_out1.clone().dtype();
            self.layernormalization61
                .forward(
                    layernormalization60_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear92_out1 = self.linear92.forward(layernormalization61_out1);
        let sigmoid41_out1 = burn::tensor::activation::sigmoid(linear92_out1.clone());
        let mul82_out1 = linear92_out1.mul(sigmoid41_out1);
        let linear93_out1 = self.linear93.forward(mul82_out1);
        let mul83_out1 = linear93_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add72_out1 = layernormalization60_out1.add(mul83_out1);
        let layernormalization62_out1 = {
            let dtype = add72_out1.clone().dtype();
            self.layernormalization62
                .forward(add72_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat31_out1 = burn::tensor::Tensor::cat(
            [gather21_out1.clone(), layernormalization62_out1.clone()].into(),
            1,
        );
        let slice45_out1 = gather21_out1.slice(s![.., 7.., ..]);
        let concat32_out1 = burn::tensor::Tensor::cat(
            [slice45_out1, layernormalization62_out1.clone()].into(),
            1,
        );
        let linear94_out1 = self.linear94.forward(layernormalization62_out1);
        let reshape72_out1 = linear94_out1.reshape([1, -1, 8, 128]);
        let linear95_out1 = self.linear95.forward(concat31_out1.clone());
        let reshape73_out1 = linear95_out1.reshape([1, -1, 8, 128]);
        let linear96_out1 = self.linear96.forward(concat31_out1);
        let reshape74_out1 = linear96_out1.reshape([1, -1, 8, 128]);
        let transpose104_out1 = reshape73_out1.permute([0, 2, 1, 3]);
        let transpose105_out1 = reshape74_out1.permute([0, 2, 1, 3]);
        let linear97_out1 = self.linear97.forward(slice2_out1);
        let reshape75_out1 = linear97_out1.reshape([1, -1, 8, 128]);
        let constant193_out1 = self.constant193.val();
        let add73_out1 = reshape72_out1
            .clone()
            .add((constant193_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose106_out1 = add73_out1.permute([0, 2, 1, 3]);
        let constant194_out1 = self.constant194.val();
        let add74_out1 = reshape72_out1
            .add((constant194_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose107_out1 = add74_out1.permute([0, 2, 1, 3]);
        let transpose108_out1 = reshape75_out1.permute([0, 2, 3, 1]);
        let matmul108_out1 = transpose107_out1.matmul(transpose108_out1);
        let pad21_out1 = matmul108_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape76_out1 = pad21_out1.reshape([1, 8, -1, 7]);
        let slice46_out1 = reshape76_out1.slice(s![.., .., 1.., ..]);
        let reshape77_out1 = slice46_out1.reshape([1, 8, 7, 153]);
        let slice47_out1 = reshape77_out1.slice(s![.., .., .., 0..77]);
        let mul84_out1 = slice47_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where31_out1 = mul84_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention11_out1,) = {
            let q = transpose106_out1;
            let k = transpose104_out1;
            let v = transpose105_out1;
            let score_bias = where31_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention11_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention11_out1 = {
                let dims = attention11_out1.dims();
                attention11_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention11_out1,)
        };
        let where32_out1 = attention11_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose109_out1 = where32_out1.permute([0, 2, 1, 3]);
        let reshape78_out1 = transpose109_out1.reshape([1, -1, 1024]);
        let linear98_out1 = self.linear98.forward(reshape78_out1);
        let add75_out1 = add72_out1.add(linear98_out1);
        let layernormalization63_out1 = {
            let dtype = add75_out1.clone().dtype();
            self.layernormalization63
                .forward(add75_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose110_out1 = layernormalization63_out1.permute([0, 2, 1]);
        let conv1d31_out1 = self.conv1d31.forward(transpose110_out1);
        let split_tensors = conv1d31_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split11_out1, split11_out2] = split_tensors.try_into().unwrap();
        let sigmoid42_out1 = burn::tensor::activation::sigmoid(split11_out2);
        let mul85_out1 = split11_out1.mul(sigmoid42_out1);
        let where33_out1 = mul85_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat33_out1 = burn::tensor::Tensor::cat(
            [gather22_out1, where33_out1].into(),
            2,
        );
        let slice48_out1 = concat33_out1.clone().slice(s![.., .., - 8..]);
        let conv1d32_out1 = self.conv1d32.forward(concat33_out1);
        let transpose111_out1 = conv1d32_out1.permute([0, 2, 1]);
        let layernormalization64_out1 = {
            let dtype = transpose111_out1.dtype();
            self.layernormalization64
                .forward(transpose111_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose112_out1 = layernormalization64_out1.permute([0, 2, 1]);
        let sigmoid43_out1 = burn::tensor::activation::sigmoid(
            transpose112_out1.clone(),
        );
        let mul86_out1 = transpose112_out1.mul(sigmoid43_out1);
        let conv1d33_out1 = self.conv1d33.forward(mul86_out1);
        let transpose113_out1 = conv1d33_out1.permute([0, 2, 1]);
        let add76_out1 = add75_out1.add(transpose113_out1);
        (add76_out1, concat32_out1, slice48_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule14 {
    layernormalization65: LayerNorm,
    linear99: Linear,
    linear100: Linear,
    layernormalization66: LayerNorm,
    layernormalization67: LayerNorm,
    linear101: Linear,
    linear102: Linear,
    layernormalization68: LayerNorm,
    linear103: Linear,
    linear104: Linear,
    linear105: Linear,
    linear106: Linear,
    constant210: burn::module::Param<Tensor<2>>,
    constant211: burn::module::Param<Tensor<2>>,
    linear107: Linear,
    layernormalization69: LayerNorm,
    conv1d34: Conv1d,
    conv1d35: Conv1d,
    layernormalization70: LayerNorm,
    conv1d36: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule14 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization65 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear99 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear100 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization66 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization67 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear101 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear102 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization68 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear103 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear104 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear105 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear106 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant210: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant211: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear107 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization69 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d34 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d35 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization70 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d36 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization65,
            linear99,
            linear100,
            layernormalization66,
            layernormalization67,
            linear101,
            linear102,
            layernormalization68,
            linear103,
            linear104,
            linear105,
            linear106,
            constant210,
            constant211,
            linear107,
            layernormalization69,
            conv1d34,
            conv1d35,
            layernormalization70,
            conv1d36,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add76_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization65_out1 = {
            let dtype = add76_out1.clone().dtype();
            self.layernormalization65
                .forward(add76_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear99_out1 = self.linear99.forward(layernormalization65_out1);
        let sigmoid44_out1 = burn::tensor::activation::sigmoid(linear99_out1.clone());
        let mul87_out1 = linear99_out1.mul(sigmoid44_out1);
        let linear100_out1 = self.linear100.forward(mul87_out1);
        let mul88_out1 = linear100_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add77_out1 = add76_out1.add(mul88_out1);
        let layernormalization66_out1 = {
            let dtype = add77_out1.dtype();
            self.layernormalization66
                .forward(add77_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather23_out1 = {
            let sliced = cache_last_channel.slice(s![11, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather24_out1 = {
            let sliced = cache_last_time.slice(s![11, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization67_out1 = {
            let dtype = layernormalization66_out1.clone().dtype();
            self.layernormalization67
                .forward(
                    layernormalization66_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear101_out1 = self.linear101.forward(layernormalization67_out1);
        let sigmoid45_out1 = burn::tensor::activation::sigmoid(linear101_out1.clone());
        let mul89_out1 = linear101_out1.mul(sigmoid45_out1);
        let linear102_out1 = self.linear102.forward(mul89_out1);
        let mul90_out1 = linear102_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add78_out1 = layernormalization66_out1.add(mul90_out1);
        let layernormalization68_out1 = {
            let dtype = add78_out1.clone().dtype();
            self.layernormalization68
                .forward(add78_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat34_out1 = burn::tensor::Tensor::cat(
            [gather23_out1.clone(), layernormalization68_out1.clone()].into(),
            1,
        );
        let slice49_out1 = gather23_out1.slice(s![.., 7.., ..]);
        let concat35_out1 = burn::tensor::Tensor::cat(
            [slice49_out1, layernormalization68_out1.clone()].into(),
            1,
        );
        let linear103_out1 = self.linear103.forward(layernormalization68_out1);
        let reshape79_out1 = linear103_out1.reshape([1, -1, 8, 128]);
        let linear104_out1 = self.linear104.forward(concat34_out1.clone());
        let reshape80_out1 = linear104_out1.reshape([1, -1, 8, 128]);
        let linear105_out1 = self.linear105.forward(concat34_out1);
        let reshape81_out1 = linear105_out1.reshape([1, -1, 8, 128]);
        let transpose114_out1 = reshape80_out1.permute([0, 2, 1, 3]);
        let transpose115_out1 = reshape81_out1.permute([0, 2, 1, 3]);
        let linear106_out1 = self.linear106.forward(slice2_out1);
        let reshape82_out1 = linear106_out1.reshape([1, -1, 8, 128]);
        let constant210_out1 = self.constant210.val();
        let add79_out1 = reshape79_out1
            .clone()
            .add((constant210_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose116_out1 = add79_out1.permute([0, 2, 1, 3]);
        let constant211_out1 = self.constant211.val();
        let add80_out1 = reshape79_out1
            .add((constant211_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose117_out1 = add80_out1.permute([0, 2, 1, 3]);
        let transpose118_out1 = reshape82_out1.permute([0, 2, 3, 1]);
        let matmul118_out1 = transpose117_out1.matmul(transpose118_out1);
        let pad23_out1 = matmul118_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape83_out1 = pad23_out1.reshape([1, 8, -1, 7]);
        let slice50_out1 = reshape83_out1.slice(s![.., .., 1.., ..]);
        let reshape84_out1 = slice50_out1.reshape([1, 8, 7, 153]);
        let slice51_out1 = reshape84_out1.slice(s![.., .., .., 0..77]);
        let mul91_out1 = slice51_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where34_out1 = mul91_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention12_out1,) = {
            let q = transpose116_out1;
            let k = transpose114_out1;
            let v = transpose115_out1;
            let score_bias = where34_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention12_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention12_out1 = {
                let dims = attention12_out1.dims();
                attention12_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention12_out1,)
        };
        let where35_out1 = attention12_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose119_out1 = where35_out1.permute([0, 2, 1, 3]);
        let reshape85_out1 = transpose119_out1.reshape([1, -1, 1024]);
        let linear107_out1 = self.linear107.forward(reshape85_out1);
        let add81_out1 = add78_out1.add(linear107_out1);
        let layernormalization69_out1 = {
            let dtype = add81_out1.clone().dtype();
            self.layernormalization69
                .forward(add81_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose120_out1 = layernormalization69_out1.permute([0, 2, 1]);
        let conv1d34_out1 = self.conv1d34.forward(transpose120_out1);
        let split_tensors = conv1d34_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split12_out1, split12_out2] = split_tensors.try_into().unwrap();
        let sigmoid46_out1 = burn::tensor::activation::sigmoid(split12_out2);
        let mul92_out1 = split12_out1.mul(sigmoid46_out1);
        let where36_out1 = mul92_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat36_out1 = burn::tensor::Tensor::cat(
            [gather24_out1, where36_out1].into(),
            2,
        );
        let slice52_out1 = concat36_out1.clone().slice(s![.., .., - 8..]);
        let conv1d35_out1 = self.conv1d35.forward(concat36_out1);
        let transpose121_out1 = conv1d35_out1.permute([0, 2, 1]);
        let layernormalization70_out1 = {
            let dtype = transpose121_out1.dtype();
            self.layernormalization70
                .forward(transpose121_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose122_out1 = layernormalization70_out1.permute([0, 2, 1]);
        let sigmoid47_out1 = burn::tensor::activation::sigmoid(
            transpose122_out1.clone(),
        );
        let mul93_out1 = transpose122_out1.mul(sigmoid47_out1);
        let conv1d36_out1 = self.conv1d36.forward(mul93_out1);
        let transpose123_out1 = conv1d36_out1.permute([0, 2, 1]);
        let add82_out1 = add81_out1.add(transpose123_out1);
        (add82_out1, concat35_out1, slice52_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule15 {
    layernormalization71: LayerNorm,
    linear108: Linear,
    linear109: Linear,
    layernormalization72: LayerNorm,
    layernormalization73: LayerNorm,
    linear110: Linear,
    linear111: Linear,
    layernormalization74: LayerNorm,
    linear112: Linear,
    linear113: Linear,
    linear114: Linear,
    linear115: Linear,
    constant227: burn::module::Param<Tensor<2>>,
    constant228: burn::module::Param<Tensor<2>>,
    linear116: Linear,
    layernormalization75: LayerNorm,
    conv1d37: Conv1d,
    conv1d38: Conv1d,
    layernormalization76: LayerNorm,
    conv1d39: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule15 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization71 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear108 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear109 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization72 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization73 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear110 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear111 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization74 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear112 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear113 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear114 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear115 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant227: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant228: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear116 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization75 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d37 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d38 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization76 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d39 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization71,
            linear108,
            linear109,
            layernormalization72,
            layernormalization73,
            linear110,
            linear111,
            layernormalization74,
            linear112,
            linear113,
            linear114,
            linear115,
            constant227,
            constant228,
            linear116,
            layernormalization75,
            conv1d37,
            conv1d38,
            layernormalization76,
            conv1d39,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add82_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization71_out1 = {
            let dtype = add82_out1.clone().dtype();
            self.layernormalization71
                .forward(add82_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear108_out1 = self.linear108.forward(layernormalization71_out1);
        let sigmoid48_out1 = burn::tensor::activation::sigmoid(linear108_out1.clone());
        let mul94_out1 = linear108_out1.mul(sigmoid48_out1);
        let linear109_out1 = self.linear109.forward(mul94_out1);
        let mul95_out1 = linear109_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add83_out1 = add82_out1.add(mul95_out1);
        let layernormalization72_out1 = {
            let dtype = add83_out1.dtype();
            self.layernormalization72
                .forward(add83_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather25_out1 = {
            let sliced = cache_last_channel.slice(s![12, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather26_out1 = {
            let sliced = cache_last_time.slice(s![12, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization73_out1 = {
            let dtype = layernormalization72_out1.clone().dtype();
            self.layernormalization73
                .forward(
                    layernormalization72_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear110_out1 = self.linear110.forward(layernormalization73_out1);
        let sigmoid49_out1 = burn::tensor::activation::sigmoid(linear110_out1.clone());
        let mul96_out1 = linear110_out1.mul(sigmoid49_out1);
        let linear111_out1 = self.linear111.forward(mul96_out1);
        let mul97_out1 = linear111_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add84_out1 = layernormalization72_out1.add(mul97_out1);
        let layernormalization74_out1 = {
            let dtype = add84_out1.clone().dtype();
            self.layernormalization74
                .forward(add84_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat37_out1 = burn::tensor::Tensor::cat(
            [gather25_out1.clone(), layernormalization74_out1.clone()].into(),
            1,
        );
        let slice53_out1 = gather25_out1.slice(s![.., 7.., ..]);
        let concat38_out1 = burn::tensor::Tensor::cat(
            [slice53_out1, layernormalization74_out1.clone()].into(),
            1,
        );
        let linear112_out1 = self.linear112.forward(layernormalization74_out1);
        let reshape86_out1 = linear112_out1.reshape([1, -1, 8, 128]);
        let linear113_out1 = self.linear113.forward(concat37_out1.clone());
        let reshape87_out1 = linear113_out1.reshape([1, -1, 8, 128]);
        let linear114_out1 = self.linear114.forward(concat37_out1);
        let reshape88_out1 = linear114_out1.reshape([1, -1, 8, 128]);
        let transpose124_out1 = reshape87_out1.permute([0, 2, 1, 3]);
        let transpose125_out1 = reshape88_out1.permute([0, 2, 1, 3]);
        let linear115_out1 = self.linear115.forward(slice2_out1);
        let reshape89_out1 = linear115_out1.reshape([1, -1, 8, 128]);
        let constant227_out1 = self.constant227.val();
        let add85_out1 = reshape86_out1
            .clone()
            .add((constant227_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose126_out1 = add85_out1.permute([0, 2, 1, 3]);
        let constant228_out1 = self.constant228.val();
        let add86_out1 = reshape86_out1
            .add((constant228_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose127_out1 = add86_out1.permute([0, 2, 1, 3]);
        let transpose128_out1 = reshape89_out1.permute([0, 2, 3, 1]);
        let matmul128_out1 = transpose127_out1.matmul(transpose128_out1);
        let pad25_out1 = matmul128_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape90_out1 = pad25_out1.reshape([1, 8, -1, 7]);
        let slice54_out1 = reshape90_out1.slice(s![.., .., 1.., ..]);
        let reshape91_out1 = slice54_out1.reshape([1, 8, 7, 153]);
        let slice55_out1 = reshape91_out1.slice(s![.., .., .., 0..77]);
        let mul98_out1 = slice55_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where37_out1 = mul98_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention13_out1,) = {
            let q = transpose126_out1;
            let k = transpose124_out1;
            let v = transpose125_out1;
            let score_bias = where37_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention13_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention13_out1 = {
                let dims = attention13_out1.dims();
                attention13_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention13_out1,)
        };
        let where38_out1 = attention13_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose129_out1 = where38_out1.permute([0, 2, 1, 3]);
        let reshape92_out1 = transpose129_out1.reshape([1, -1, 1024]);
        let linear116_out1 = self.linear116.forward(reshape92_out1);
        let add87_out1 = add84_out1.add(linear116_out1);
        let layernormalization75_out1 = {
            let dtype = add87_out1.clone().dtype();
            self.layernormalization75
                .forward(add87_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose130_out1 = layernormalization75_out1.permute([0, 2, 1]);
        let conv1d37_out1 = self.conv1d37.forward(transpose130_out1);
        let split_tensors = conv1d37_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split13_out1, split13_out2] = split_tensors.try_into().unwrap();
        let sigmoid50_out1 = burn::tensor::activation::sigmoid(split13_out2);
        let mul99_out1 = split13_out1.mul(sigmoid50_out1);
        let where39_out1 = mul99_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat39_out1 = burn::tensor::Tensor::cat(
            [gather26_out1, where39_out1].into(),
            2,
        );
        let slice56_out1 = concat39_out1.clone().slice(s![.., .., - 8..]);
        let conv1d38_out1 = self.conv1d38.forward(concat39_out1);
        let transpose131_out1 = conv1d38_out1.permute([0, 2, 1]);
        let layernormalization76_out1 = {
            let dtype = transpose131_out1.dtype();
            self.layernormalization76
                .forward(transpose131_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose132_out1 = layernormalization76_out1.permute([0, 2, 1]);
        let sigmoid51_out1 = burn::tensor::activation::sigmoid(
            transpose132_out1.clone(),
        );
        let mul100_out1 = transpose132_out1.mul(sigmoid51_out1);
        let conv1d39_out1 = self.conv1d39.forward(mul100_out1);
        let transpose133_out1 = conv1d39_out1.permute([0, 2, 1]);
        let add88_out1 = add87_out1.add(transpose133_out1);
        (add88_out1, concat38_out1, slice56_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule16 {
    layernormalization77: LayerNorm,
    linear117: Linear,
    linear118: Linear,
    layernormalization78: LayerNorm,
    layernormalization79: LayerNorm,
    linear119: Linear,
    linear120: Linear,
    layernormalization80: LayerNorm,
    linear121: Linear,
    linear122: Linear,
    linear123: Linear,
    linear124: Linear,
    constant244: burn::module::Param<Tensor<2>>,
    constant245: burn::module::Param<Tensor<2>>,
    linear125: Linear,
    layernormalization81: LayerNorm,
    conv1d40: Conv1d,
    conv1d41: Conv1d,
    layernormalization82: LayerNorm,
    conv1d42: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule16 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization77 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear117 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear118 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization78 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization79 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear119 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear120 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization80 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear121 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear122 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear123 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear124 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant244: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant245: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear125 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization81 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d40 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d41 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization82 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d42 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization77,
            linear117,
            linear118,
            layernormalization78,
            layernormalization79,
            linear119,
            linear120,
            layernormalization80,
            linear121,
            linear122,
            linear123,
            linear124,
            constant244,
            constant245,
            linear125,
            layernormalization81,
            conv1d40,
            conv1d41,
            layernormalization82,
            conv1d42,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add88_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization77_out1 = {
            let dtype = add88_out1.clone().dtype();
            self.layernormalization77
                .forward(add88_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear117_out1 = self.linear117.forward(layernormalization77_out1);
        let sigmoid52_out1 = burn::tensor::activation::sigmoid(linear117_out1.clone());
        let mul101_out1 = linear117_out1.mul(sigmoid52_out1);
        let linear118_out1 = self.linear118.forward(mul101_out1);
        let mul102_out1 = linear118_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add89_out1 = add88_out1.add(mul102_out1);
        let layernormalization78_out1 = {
            let dtype = add89_out1.dtype();
            self.layernormalization78
                .forward(add89_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather27_out1 = {
            let sliced = cache_last_channel.slice(s![13, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather28_out1 = {
            let sliced = cache_last_time.slice(s![13, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization79_out1 = {
            let dtype = layernormalization78_out1.clone().dtype();
            self.layernormalization79
                .forward(
                    layernormalization78_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear119_out1 = self.linear119.forward(layernormalization79_out1);
        let sigmoid53_out1 = burn::tensor::activation::sigmoid(linear119_out1.clone());
        let mul103_out1 = linear119_out1.mul(sigmoid53_out1);
        let linear120_out1 = self.linear120.forward(mul103_out1);
        let mul104_out1 = linear120_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add90_out1 = layernormalization78_out1.add(mul104_out1);
        let layernormalization80_out1 = {
            let dtype = add90_out1.clone().dtype();
            self.layernormalization80
                .forward(add90_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat40_out1 = burn::tensor::Tensor::cat(
            [gather27_out1.clone(), layernormalization80_out1.clone()].into(),
            1,
        );
        let slice57_out1 = gather27_out1.slice(s![.., 7.., ..]);
        let concat41_out1 = burn::tensor::Tensor::cat(
            [slice57_out1, layernormalization80_out1.clone()].into(),
            1,
        );
        let linear121_out1 = self.linear121.forward(layernormalization80_out1);
        let reshape93_out1 = linear121_out1.reshape([1, -1, 8, 128]);
        let linear122_out1 = self.linear122.forward(concat40_out1.clone());
        let reshape94_out1 = linear122_out1.reshape([1, -1, 8, 128]);
        let linear123_out1 = self.linear123.forward(concat40_out1);
        let reshape95_out1 = linear123_out1.reshape([1, -1, 8, 128]);
        let transpose134_out1 = reshape94_out1.permute([0, 2, 1, 3]);
        let transpose135_out1 = reshape95_out1.permute([0, 2, 1, 3]);
        let linear124_out1 = self.linear124.forward(slice2_out1);
        let reshape96_out1 = linear124_out1.reshape([1, -1, 8, 128]);
        let constant244_out1 = self.constant244.val();
        let add91_out1 = reshape93_out1
            .clone()
            .add((constant244_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose136_out1 = add91_out1.permute([0, 2, 1, 3]);
        let constant245_out1 = self.constant245.val();
        let add92_out1 = reshape93_out1
            .add((constant245_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose137_out1 = add92_out1.permute([0, 2, 1, 3]);
        let transpose138_out1 = reshape96_out1.permute([0, 2, 3, 1]);
        let matmul138_out1 = transpose137_out1.matmul(transpose138_out1);
        let pad27_out1 = matmul138_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape97_out1 = pad27_out1.reshape([1, 8, -1, 7]);
        let slice58_out1 = reshape97_out1.slice(s![.., .., 1.., ..]);
        let reshape98_out1 = slice58_out1.reshape([1, 8, 7, 153]);
        let slice59_out1 = reshape98_out1.slice(s![.., .., .., 0..77]);
        let mul105_out1 = slice59_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where40_out1 = mul105_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention14_out1,) = {
            let q = transpose136_out1;
            let k = transpose134_out1;
            let v = transpose135_out1;
            let score_bias = where40_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention14_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention14_out1 = {
                let dims = attention14_out1.dims();
                attention14_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention14_out1,)
        };
        let where41_out1 = attention14_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose139_out1 = where41_out1.permute([0, 2, 1, 3]);
        let reshape99_out1 = transpose139_out1.reshape([1, -1, 1024]);
        let linear125_out1 = self.linear125.forward(reshape99_out1);
        let add93_out1 = add90_out1.add(linear125_out1);
        let layernormalization81_out1 = {
            let dtype = add93_out1.clone().dtype();
            self.layernormalization81
                .forward(add93_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose140_out1 = layernormalization81_out1.permute([0, 2, 1]);
        let conv1d40_out1 = self.conv1d40.forward(transpose140_out1);
        let split_tensors = conv1d40_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split14_out1, split14_out2] = split_tensors.try_into().unwrap();
        let sigmoid54_out1 = burn::tensor::activation::sigmoid(split14_out2);
        let mul106_out1 = split14_out1.mul(sigmoid54_out1);
        let where42_out1 = mul106_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat42_out1 = burn::tensor::Tensor::cat(
            [gather28_out1, where42_out1].into(),
            2,
        );
        let slice60_out1 = concat42_out1.clone().slice(s![.., .., - 8..]);
        let conv1d41_out1 = self.conv1d41.forward(concat42_out1);
        let transpose141_out1 = conv1d41_out1.permute([0, 2, 1]);
        let layernormalization82_out1 = {
            let dtype = transpose141_out1.dtype();
            self.layernormalization82
                .forward(transpose141_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose142_out1 = layernormalization82_out1.permute([0, 2, 1]);
        let sigmoid55_out1 = burn::tensor::activation::sigmoid(
            transpose142_out1.clone(),
        );
        let mul107_out1 = transpose142_out1.mul(sigmoid55_out1);
        let conv1d42_out1 = self.conv1d42.forward(mul107_out1);
        let transpose143_out1 = conv1d42_out1.permute([0, 2, 1]);
        let add94_out1 = add93_out1.add(transpose143_out1);
        (add94_out1, concat41_out1, slice60_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule17 {
    layernormalization83: LayerNorm,
    linear126: Linear,
    linear127: Linear,
    layernormalization84: LayerNorm,
    layernormalization85: LayerNorm,
    linear128: Linear,
    linear129: Linear,
    layernormalization86: LayerNorm,
    linear130: Linear,
    linear131: Linear,
    linear132: Linear,
    linear133: Linear,
    constant261: burn::module::Param<Tensor<2>>,
    constant262: burn::module::Param<Tensor<2>>,
    linear134: Linear,
    layernormalization87: LayerNorm,
    conv1d43: Conv1d,
    conv1d44: Conv1d,
    layernormalization88: LayerNorm,
    conv1d45: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule17 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization83 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear126 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear127 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization84 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization85 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear128 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear129 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization86 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear130 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear131 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear132 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear133 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant261: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant262: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear134 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization87 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d43 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d44 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization88 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d45 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization83,
            linear126,
            linear127,
            layernormalization84,
            layernormalization85,
            linear128,
            linear129,
            layernormalization86,
            linear130,
            linear131,
            linear132,
            linear133,
            constant261,
            constant262,
            linear134,
            layernormalization87,
            conv1d43,
            conv1d44,
            layernormalization88,
            conv1d45,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add94_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization83_out1 = {
            let dtype = add94_out1.clone().dtype();
            self.layernormalization83
                .forward(add94_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear126_out1 = self.linear126.forward(layernormalization83_out1);
        let sigmoid56_out1 = burn::tensor::activation::sigmoid(linear126_out1.clone());
        let mul108_out1 = linear126_out1.mul(sigmoid56_out1);
        let linear127_out1 = self.linear127.forward(mul108_out1);
        let mul109_out1 = linear127_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add95_out1 = add94_out1.add(mul109_out1);
        let layernormalization84_out1 = {
            let dtype = add95_out1.dtype();
            self.layernormalization84
                .forward(add95_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather29_out1 = {
            let sliced = cache_last_channel.slice(s![14, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather30_out1 = {
            let sliced = cache_last_time.slice(s![14, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization85_out1 = {
            let dtype = layernormalization84_out1.clone().dtype();
            self.layernormalization85
                .forward(
                    layernormalization84_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear128_out1 = self.linear128.forward(layernormalization85_out1);
        let sigmoid57_out1 = burn::tensor::activation::sigmoid(linear128_out1.clone());
        let mul110_out1 = linear128_out1.mul(sigmoid57_out1);
        let linear129_out1 = self.linear129.forward(mul110_out1);
        let mul111_out1 = linear129_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add96_out1 = layernormalization84_out1.add(mul111_out1);
        let layernormalization86_out1 = {
            let dtype = add96_out1.clone().dtype();
            self.layernormalization86
                .forward(add96_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat43_out1 = burn::tensor::Tensor::cat(
            [gather29_out1.clone(), layernormalization86_out1.clone()].into(),
            1,
        );
        let slice61_out1 = gather29_out1.slice(s![.., 7.., ..]);
        let concat44_out1 = burn::tensor::Tensor::cat(
            [slice61_out1, layernormalization86_out1.clone()].into(),
            1,
        );
        let linear130_out1 = self.linear130.forward(layernormalization86_out1);
        let reshape100_out1 = linear130_out1.reshape([1, -1, 8, 128]);
        let linear131_out1 = self.linear131.forward(concat43_out1.clone());
        let reshape101_out1 = linear131_out1.reshape([1, -1, 8, 128]);
        let linear132_out1 = self.linear132.forward(concat43_out1);
        let reshape102_out1 = linear132_out1.reshape([1, -1, 8, 128]);
        let transpose144_out1 = reshape101_out1.permute([0, 2, 1, 3]);
        let transpose145_out1 = reshape102_out1.permute([0, 2, 1, 3]);
        let linear133_out1 = self.linear133.forward(slice2_out1);
        let reshape103_out1 = linear133_out1.reshape([1, -1, 8, 128]);
        let constant261_out1 = self.constant261.val();
        let add97_out1 = reshape100_out1
            .clone()
            .add((constant261_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose146_out1 = add97_out1.permute([0, 2, 1, 3]);
        let constant262_out1 = self.constant262.val();
        let add98_out1 = reshape100_out1
            .add((constant262_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose147_out1 = add98_out1.permute([0, 2, 1, 3]);
        let transpose148_out1 = reshape103_out1.permute([0, 2, 3, 1]);
        let matmul148_out1 = transpose147_out1.matmul(transpose148_out1);
        let pad29_out1 = matmul148_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape104_out1 = pad29_out1.reshape([1, 8, -1, 7]);
        let slice62_out1 = reshape104_out1.slice(s![.., .., 1.., ..]);
        let reshape105_out1 = slice62_out1.reshape([1, 8, 7, 153]);
        let slice63_out1 = reshape105_out1.slice(s![.., .., .., 0..77]);
        let mul112_out1 = slice63_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where43_out1 = mul112_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention15_out1,) = {
            let q = transpose146_out1;
            let k = transpose144_out1;
            let v = transpose145_out1;
            let score_bias = where43_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention15_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention15_out1 = {
                let dims = attention15_out1.dims();
                attention15_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention15_out1,)
        };
        let where44_out1 = attention15_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose149_out1 = where44_out1.permute([0, 2, 1, 3]);
        let reshape106_out1 = transpose149_out1.reshape([1, -1, 1024]);
        let linear134_out1 = self.linear134.forward(reshape106_out1);
        let add99_out1 = add96_out1.add(linear134_out1);
        let layernormalization87_out1 = {
            let dtype = add99_out1.clone().dtype();
            self.layernormalization87
                .forward(add99_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose150_out1 = layernormalization87_out1.permute([0, 2, 1]);
        let conv1d43_out1 = self.conv1d43.forward(transpose150_out1);
        let split_tensors = conv1d43_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split15_out1, split15_out2] = split_tensors.try_into().unwrap();
        let sigmoid58_out1 = burn::tensor::activation::sigmoid(split15_out2);
        let mul113_out1 = split15_out1.mul(sigmoid58_out1);
        let where45_out1 = mul113_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat45_out1 = burn::tensor::Tensor::cat(
            [gather30_out1, where45_out1].into(),
            2,
        );
        let slice64_out1 = concat45_out1.clone().slice(s![.., .., - 8..]);
        let conv1d44_out1 = self.conv1d44.forward(concat45_out1);
        let transpose151_out1 = conv1d44_out1.permute([0, 2, 1]);
        let layernormalization88_out1 = {
            let dtype = transpose151_out1.dtype();
            self.layernormalization88
                .forward(transpose151_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose152_out1 = layernormalization88_out1.permute([0, 2, 1]);
        let sigmoid59_out1 = burn::tensor::activation::sigmoid(
            transpose152_out1.clone(),
        );
        let mul114_out1 = transpose152_out1.mul(sigmoid59_out1);
        let conv1d45_out1 = self.conv1d45.forward(mul114_out1);
        let transpose153_out1 = conv1d45_out1.permute([0, 2, 1]);
        let add100_out1 = add99_out1.add(transpose153_out1);
        (add100_out1, concat44_out1, slice64_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule18 {
    layernormalization89: LayerNorm,
    linear135: Linear,
    linear136: Linear,
    layernormalization90: LayerNorm,
    layernormalization91: LayerNorm,
    linear137: Linear,
    linear138: Linear,
    layernormalization92: LayerNorm,
    linear139: Linear,
    linear140: Linear,
    linear141: Linear,
    linear142: Linear,
    constant278: burn::module::Param<Tensor<2>>,
    constant279: burn::module::Param<Tensor<2>>,
    linear143: Linear,
    layernormalization93: LayerNorm,
    conv1d46: Conv1d,
    conv1d47: Conv1d,
    layernormalization94: LayerNorm,
    conv1d48: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule18 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization89 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear135 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear136 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization90 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization91 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear137 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear138 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization92 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear139 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear140 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear141 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear142 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant278: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant279: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear143 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization93 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d46 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d47 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization94 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d48 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization89,
            linear135,
            linear136,
            layernormalization90,
            layernormalization91,
            linear137,
            linear138,
            layernormalization92,
            linear139,
            linear140,
            linear141,
            linear142,
            constant278,
            constant279,
            linear143,
            layernormalization93,
            conv1d46,
            conv1d47,
            layernormalization94,
            conv1d48,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add100_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization89_out1 = {
            let dtype = add100_out1.clone().dtype();
            self.layernormalization89
                .forward(add100_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear135_out1 = self.linear135.forward(layernormalization89_out1);
        let sigmoid60_out1 = burn::tensor::activation::sigmoid(linear135_out1.clone());
        let mul115_out1 = linear135_out1.mul(sigmoid60_out1);
        let linear136_out1 = self.linear136.forward(mul115_out1);
        let mul116_out1 = linear136_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add101_out1 = add100_out1.add(mul116_out1);
        let layernormalization90_out1 = {
            let dtype = add101_out1.dtype();
            self.layernormalization90
                .forward(add101_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather31_out1 = {
            let sliced = cache_last_channel.slice(s![15, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather32_out1 = {
            let sliced = cache_last_time.slice(s![15, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization91_out1 = {
            let dtype = layernormalization90_out1.clone().dtype();
            self.layernormalization91
                .forward(
                    layernormalization90_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear137_out1 = self.linear137.forward(layernormalization91_out1);
        let sigmoid61_out1 = burn::tensor::activation::sigmoid(linear137_out1.clone());
        let mul117_out1 = linear137_out1.mul(sigmoid61_out1);
        let linear138_out1 = self.linear138.forward(mul117_out1);
        let mul118_out1 = linear138_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add102_out1 = layernormalization90_out1.add(mul118_out1);
        let layernormalization92_out1 = {
            let dtype = add102_out1.clone().dtype();
            self.layernormalization92
                .forward(add102_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat46_out1 = burn::tensor::Tensor::cat(
            [gather31_out1.clone(), layernormalization92_out1.clone()].into(),
            1,
        );
        let slice65_out1 = gather31_out1.slice(s![.., 7.., ..]);
        let concat47_out1 = burn::tensor::Tensor::cat(
            [slice65_out1, layernormalization92_out1.clone()].into(),
            1,
        );
        let linear139_out1 = self.linear139.forward(layernormalization92_out1);
        let reshape107_out1 = linear139_out1.reshape([1, -1, 8, 128]);
        let linear140_out1 = self.linear140.forward(concat46_out1.clone());
        let reshape108_out1 = linear140_out1.reshape([1, -1, 8, 128]);
        let linear141_out1 = self.linear141.forward(concat46_out1);
        let reshape109_out1 = linear141_out1.reshape([1, -1, 8, 128]);
        let transpose154_out1 = reshape108_out1.permute([0, 2, 1, 3]);
        let transpose155_out1 = reshape109_out1.permute([0, 2, 1, 3]);
        let linear142_out1 = self.linear142.forward(slice2_out1);
        let reshape110_out1 = linear142_out1.reshape([1, -1, 8, 128]);
        let constant278_out1 = self.constant278.val();
        let add103_out1 = reshape107_out1
            .clone()
            .add((constant278_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose156_out1 = add103_out1.permute([0, 2, 1, 3]);
        let constant279_out1 = self.constant279.val();
        let add104_out1 = reshape107_out1
            .add((constant279_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose157_out1 = add104_out1.permute([0, 2, 1, 3]);
        let transpose158_out1 = reshape110_out1.permute([0, 2, 3, 1]);
        let matmul158_out1 = transpose157_out1.matmul(transpose158_out1);
        let pad31_out1 = matmul158_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape111_out1 = pad31_out1.reshape([1, 8, -1, 7]);
        let slice66_out1 = reshape111_out1.slice(s![.., .., 1.., ..]);
        let reshape112_out1 = slice66_out1.reshape([1, 8, 7, 153]);
        let slice67_out1 = reshape112_out1.slice(s![.., .., .., 0..77]);
        let mul119_out1 = slice67_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where46_out1 = mul119_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention16_out1,) = {
            let q = transpose156_out1;
            let k = transpose154_out1;
            let v = transpose155_out1;
            let score_bias = where46_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention16_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention16_out1 = {
                let dims = attention16_out1.dims();
                attention16_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention16_out1,)
        };
        let where47_out1 = attention16_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose159_out1 = where47_out1.permute([0, 2, 1, 3]);
        let reshape113_out1 = transpose159_out1.reshape([1, -1, 1024]);
        let linear143_out1 = self.linear143.forward(reshape113_out1);
        let add105_out1 = add102_out1.add(linear143_out1);
        let layernormalization93_out1 = {
            let dtype = add105_out1.clone().dtype();
            self.layernormalization93
                .forward(add105_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose160_out1 = layernormalization93_out1.permute([0, 2, 1]);
        let conv1d46_out1 = self.conv1d46.forward(transpose160_out1);
        let split_tensors = conv1d46_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split16_out1, split16_out2] = split_tensors.try_into().unwrap();
        let sigmoid62_out1 = burn::tensor::activation::sigmoid(split16_out2);
        let mul120_out1 = split16_out1.mul(sigmoid62_out1);
        let where48_out1 = mul120_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat48_out1 = burn::tensor::Tensor::cat(
            [gather32_out1, where48_out1].into(),
            2,
        );
        let slice68_out1 = concat48_out1.clone().slice(s![.., .., - 8..]);
        let conv1d47_out1 = self.conv1d47.forward(concat48_out1);
        let transpose161_out1 = conv1d47_out1.permute([0, 2, 1]);
        let layernormalization94_out1 = {
            let dtype = transpose161_out1.dtype();
            self.layernormalization94
                .forward(transpose161_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose162_out1 = layernormalization94_out1.permute([0, 2, 1]);
        let sigmoid63_out1 = burn::tensor::activation::sigmoid(
            transpose162_out1.clone(),
        );
        let mul121_out1 = transpose162_out1.mul(sigmoid63_out1);
        let conv1d48_out1 = self.conv1d48.forward(mul121_out1);
        let transpose163_out1 = conv1d48_out1.permute([0, 2, 1]);
        let add106_out1 = add105_out1.add(transpose163_out1);
        (add106_out1, concat47_out1, slice68_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule19 {
    layernormalization95: LayerNorm,
    linear144: Linear,
    linear145: Linear,
    layernormalization96: LayerNorm,
    layernormalization97: LayerNorm,
    linear146: Linear,
    linear147: Linear,
    layernormalization98: LayerNorm,
    linear148: Linear,
    linear149: Linear,
    linear150: Linear,
    linear151: Linear,
    constant295: burn::module::Param<Tensor<2>>,
    constant296: burn::module::Param<Tensor<2>>,
    linear152: Linear,
    layernormalization99: LayerNorm,
    conv1d49: Conv1d,
    conv1d50: Conv1d,
    layernormalization100: LayerNorm,
    conv1d51: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule19 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization95 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear144 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear145 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization96 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization97 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear146 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear147 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization98 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear148 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear149 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear150 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear151 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant295: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant296: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear152 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization99 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d49 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d50 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization100 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d51 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization95,
            linear144,
            linear145,
            layernormalization96,
            layernormalization97,
            linear146,
            linear147,
            layernormalization98,
            linear148,
            linear149,
            linear150,
            linear151,
            constant295,
            constant296,
            linear152,
            layernormalization99,
            conv1d49,
            conv1d50,
            layernormalization100,
            conv1d51,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add106_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization95_out1 = {
            let dtype = add106_out1.clone().dtype();
            self.layernormalization95
                .forward(add106_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear144_out1 = self.linear144.forward(layernormalization95_out1);
        let sigmoid64_out1 = burn::tensor::activation::sigmoid(linear144_out1.clone());
        let mul122_out1 = linear144_out1.mul(sigmoid64_out1);
        let linear145_out1 = self.linear145.forward(mul122_out1);
        let mul123_out1 = linear145_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add107_out1 = add106_out1.add(mul123_out1);
        let layernormalization96_out1 = {
            let dtype = add107_out1.dtype();
            self.layernormalization96
                .forward(add107_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather33_out1 = {
            let sliced = cache_last_channel.slice(s![16, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather34_out1 = {
            let sliced = cache_last_time.slice(s![16, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization97_out1 = {
            let dtype = layernormalization96_out1.clone().dtype();
            self.layernormalization97
                .forward(
                    layernormalization96_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear146_out1 = self.linear146.forward(layernormalization97_out1);
        let sigmoid65_out1 = burn::tensor::activation::sigmoid(linear146_out1.clone());
        let mul124_out1 = linear146_out1.mul(sigmoid65_out1);
        let linear147_out1 = self.linear147.forward(mul124_out1);
        let mul125_out1 = linear147_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add108_out1 = layernormalization96_out1.add(mul125_out1);
        let layernormalization98_out1 = {
            let dtype = add108_out1.clone().dtype();
            self.layernormalization98
                .forward(add108_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat49_out1 = burn::tensor::Tensor::cat(
            [gather33_out1.clone(), layernormalization98_out1.clone()].into(),
            1,
        );
        let slice69_out1 = gather33_out1.slice(s![.., 7.., ..]);
        let concat50_out1 = burn::tensor::Tensor::cat(
            [slice69_out1, layernormalization98_out1.clone()].into(),
            1,
        );
        let linear148_out1 = self.linear148.forward(layernormalization98_out1);
        let reshape114_out1 = linear148_out1.reshape([1, -1, 8, 128]);
        let linear149_out1 = self.linear149.forward(concat49_out1.clone());
        let reshape115_out1 = linear149_out1.reshape([1, -1, 8, 128]);
        let linear150_out1 = self.linear150.forward(concat49_out1);
        let reshape116_out1 = linear150_out1.reshape([1, -1, 8, 128]);
        let transpose164_out1 = reshape115_out1.permute([0, 2, 1, 3]);
        let transpose165_out1 = reshape116_out1.permute([0, 2, 1, 3]);
        let linear151_out1 = self.linear151.forward(slice2_out1);
        let reshape117_out1 = linear151_out1.reshape([1, -1, 8, 128]);
        let constant295_out1 = self.constant295.val();
        let add109_out1 = reshape114_out1
            .clone()
            .add((constant295_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose166_out1 = add109_out1.permute([0, 2, 1, 3]);
        let constant296_out1 = self.constant296.val();
        let add110_out1 = reshape114_out1
            .add((constant296_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose167_out1 = add110_out1.permute([0, 2, 1, 3]);
        let transpose168_out1 = reshape117_out1.permute([0, 2, 3, 1]);
        let matmul168_out1 = transpose167_out1.matmul(transpose168_out1);
        let pad33_out1 = matmul168_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape118_out1 = pad33_out1.reshape([1, 8, -1, 7]);
        let slice70_out1 = reshape118_out1.slice(s![.., .., 1.., ..]);
        let reshape119_out1 = slice70_out1.reshape([1, 8, 7, 153]);
        let slice71_out1 = reshape119_out1.slice(s![.., .., .., 0..77]);
        let mul126_out1 = slice71_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where49_out1 = mul126_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention17_out1,) = {
            let q = transpose166_out1;
            let k = transpose164_out1;
            let v = transpose165_out1;
            let score_bias = where49_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention17_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention17_out1 = {
                let dims = attention17_out1.dims();
                attention17_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention17_out1,)
        };
        let where50_out1 = attention17_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose169_out1 = where50_out1.permute([0, 2, 1, 3]);
        let reshape120_out1 = transpose169_out1.reshape([1, -1, 1024]);
        let linear152_out1 = self.linear152.forward(reshape120_out1);
        let add111_out1 = add108_out1.add(linear152_out1);
        let layernormalization99_out1 = {
            let dtype = add111_out1.clone().dtype();
            self.layernormalization99
                .forward(add111_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose170_out1 = layernormalization99_out1.permute([0, 2, 1]);
        let conv1d49_out1 = self.conv1d49.forward(transpose170_out1);
        let split_tensors = conv1d49_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split17_out1, split17_out2] = split_tensors.try_into().unwrap();
        let sigmoid66_out1 = burn::tensor::activation::sigmoid(split17_out2);
        let mul127_out1 = split17_out1.mul(sigmoid66_out1);
        let where51_out1 = mul127_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat51_out1 = burn::tensor::Tensor::cat(
            [gather34_out1, where51_out1].into(),
            2,
        );
        let slice72_out1 = concat51_out1.clone().slice(s![.., .., - 8..]);
        let conv1d50_out1 = self.conv1d50.forward(concat51_out1);
        let transpose171_out1 = conv1d50_out1.permute([0, 2, 1]);
        let layernormalization100_out1 = {
            let dtype = transpose171_out1.dtype();
            self.layernormalization100
                .forward(transpose171_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose172_out1 = layernormalization100_out1.permute([0, 2, 1]);
        let sigmoid67_out1 = burn::tensor::activation::sigmoid(
            transpose172_out1.clone(),
        );
        let mul128_out1 = transpose172_out1.mul(sigmoid67_out1);
        let conv1d51_out1 = self.conv1d51.forward(mul128_out1);
        let transpose173_out1 = conv1d51_out1.permute([0, 2, 1]);
        let add112_out1 = add111_out1.add(transpose173_out1);
        (add112_out1, concat50_out1, slice72_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule20 {
    layernormalization101: LayerNorm,
    linear153: Linear,
    linear154: Linear,
    layernormalization102: LayerNorm,
    layernormalization103: LayerNorm,
    linear155: Linear,
    linear156: Linear,
    layernormalization104: LayerNorm,
    linear157: Linear,
    linear158: Linear,
    linear159: Linear,
    linear160: Linear,
    constant312: burn::module::Param<Tensor<2>>,
    constant313: burn::module::Param<Tensor<2>>,
    linear161: Linear,
    layernormalization105: LayerNorm,
    conv1d52: Conv1d,
    conv1d53: Conv1d,
    layernormalization106: LayerNorm,
    conv1d54: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule20 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization101 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear153 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear154 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization102 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization103 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear155 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear156 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization104 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear157 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear158 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear159 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear160 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant312: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant313: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear161 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization105 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d52 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d53 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization106 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d54 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization101,
            linear153,
            linear154,
            layernormalization102,
            layernormalization103,
            linear155,
            linear156,
            layernormalization104,
            linear157,
            linear158,
            linear159,
            linear160,
            constant312,
            constant313,
            linear161,
            layernormalization105,
            conv1d52,
            conv1d53,
            layernormalization106,
            conv1d54,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add112_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization101_out1 = {
            let dtype = add112_out1.clone().dtype();
            self.layernormalization101
                .forward(add112_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear153_out1 = self.linear153.forward(layernormalization101_out1);
        let sigmoid68_out1 = burn::tensor::activation::sigmoid(linear153_out1.clone());
        let mul129_out1 = linear153_out1.mul(sigmoid68_out1);
        let linear154_out1 = self.linear154.forward(mul129_out1);
        let mul130_out1 = linear154_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add113_out1 = add112_out1.add(mul130_out1);
        let layernormalization102_out1 = {
            let dtype = add113_out1.dtype();
            self.layernormalization102
                .forward(add113_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather35_out1 = {
            let sliced = cache_last_channel.slice(s![17, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather36_out1 = {
            let sliced = cache_last_time.slice(s![17, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization103_out1 = {
            let dtype = layernormalization102_out1.clone().dtype();
            self.layernormalization103
                .forward(
                    layernormalization102_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear155_out1 = self.linear155.forward(layernormalization103_out1);
        let sigmoid69_out1 = burn::tensor::activation::sigmoid(linear155_out1.clone());
        let mul131_out1 = linear155_out1.mul(sigmoid69_out1);
        let linear156_out1 = self.linear156.forward(mul131_out1);
        let mul132_out1 = linear156_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add114_out1 = layernormalization102_out1.add(mul132_out1);
        let layernormalization104_out1 = {
            let dtype = add114_out1.clone().dtype();
            self.layernormalization104
                .forward(add114_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat52_out1 = burn::tensor::Tensor::cat(
            [gather35_out1.clone(), layernormalization104_out1.clone()].into(),
            1,
        );
        let slice73_out1 = gather35_out1.slice(s![.., 7.., ..]);
        let concat53_out1 = burn::tensor::Tensor::cat(
            [slice73_out1, layernormalization104_out1.clone()].into(),
            1,
        );
        let linear157_out1 = self.linear157.forward(layernormalization104_out1);
        let reshape121_out1 = linear157_out1.reshape([1, -1, 8, 128]);
        let linear158_out1 = self.linear158.forward(concat52_out1.clone());
        let reshape122_out1 = linear158_out1.reshape([1, -1, 8, 128]);
        let linear159_out1 = self.linear159.forward(concat52_out1);
        let reshape123_out1 = linear159_out1.reshape([1, -1, 8, 128]);
        let transpose174_out1 = reshape122_out1.permute([0, 2, 1, 3]);
        let transpose175_out1 = reshape123_out1.permute([0, 2, 1, 3]);
        let linear160_out1 = self.linear160.forward(slice2_out1);
        let reshape124_out1 = linear160_out1.reshape([1, -1, 8, 128]);
        let constant312_out1 = self.constant312.val();
        let add115_out1 = reshape121_out1
            .clone()
            .add((constant312_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose176_out1 = add115_out1.permute([0, 2, 1, 3]);
        let constant313_out1 = self.constant313.val();
        let add116_out1 = reshape121_out1
            .add((constant313_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose177_out1 = add116_out1.permute([0, 2, 1, 3]);
        let transpose178_out1 = reshape124_out1.permute([0, 2, 3, 1]);
        let matmul178_out1 = transpose177_out1.matmul(transpose178_out1);
        let pad35_out1 = matmul178_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape125_out1 = pad35_out1.reshape([1, 8, -1, 7]);
        let slice74_out1 = reshape125_out1.slice(s![.., .., 1.., ..]);
        let reshape126_out1 = slice74_out1.reshape([1, 8, 7, 153]);
        let slice75_out1 = reshape126_out1.slice(s![.., .., .., 0..77]);
        let mul133_out1 = slice75_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where52_out1 = mul133_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention18_out1,) = {
            let q = transpose176_out1;
            let k = transpose174_out1;
            let v = transpose175_out1;
            let score_bias = where52_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention18_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention18_out1 = {
                let dims = attention18_out1.dims();
                attention18_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention18_out1,)
        };
        let where53_out1 = attention18_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose179_out1 = where53_out1.permute([0, 2, 1, 3]);
        let reshape127_out1 = transpose179_out1.reshape([1, -1, 1024]);
        let linear161_out1 = self.linear161.forward(reshape127_out1);
        let add117_out1 = add114_out1.add(linear161_out1);
        let layernormalization105_out1 = {
            let dtype = add117_out1.clone().dtype();
            self.layernormalization105
                .forward(add117_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose180_out1 = layernormalization105_out1.permute([0, 2, 1]);
        let conv1d52_out1 = self.conv1d52.forward(transpose180_out1);
        let split_tensors = conv1d52_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split18_out1, split18_out2] = split_tensors.try_into().unwrap();
        let sigmoid70_out1 = burn::tensor::activation::sigmoid(split18_out2);
        let mul134_out1 = split18_out1.mul(sigmoid70_out1);
        let where54_out1 = mul134_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat54_out1 = burn::tensor::Tensor::cat(
            [gather36_out1, where54_out1].into(),
            2,
        );
        let slice76_out1 = concat54_out1.clone().slice(s![.., .., - 8..]);
        let conv1d53_out1 = self.conv1d53.forward(concat54_out1);
        let transpose181_out1 = conv1d53_out1.permute([0, 2, 1]);
        let layernormalization106_out1 = {
            let dtype = transpose181_out1.dtype();
            self.layernormalization106
                .forward(transpose181_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose182_out1 = layernormalization106_out1.permute([0, 2, 1]);
        let sigmoid71_out1 = burn::tensor::activation::sigmoid(
            transpose182_out1.clone(),
        );
        let mul135_out1 = transpose182_out1.mul(sigmoid71_out1);
        let conv1d54_out1 = self.conv1d54.forward(mul135_out1);
        let transpose183_out1 = conv1d54_out1.permute([0, 2, 1]);
        let add118_out1 = add117_out1.add(transpose183_out1);
        (add118_out1, concat53_out1, slice76_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule21 {
    layernormalization107: LayerNorm,
    linear162: Linear,
    linear163: Linear,
    layernormalization108: LayerNorm,
    layernormalization109: LayerNorm,
    linear164: Linear,
    linear165: Linear,
    layernormalization110: LayerNorm,
    linear166: Linear,
    linear167: Linear,
    linear168: Linear,
    linear169: Linear,
    constant329: burn::module::Param<Tensor<2>>,
    constant330: burn::module::Param<Tensor<2>>,
    linear170: Linear,
    layernormalization111: LayerNorm,
    conv1d55: Conv1d,
    conv1d56: Conv1d,
    layernormalization112: LayerNorm,
    conv1d57: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule21 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization107 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear162 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear163 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization108 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization109 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear164 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear165 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization110 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear166 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear167 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear168 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear169 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant329: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant330: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear170 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization111 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d55 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d56 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization112 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d57 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization107,
            linear162,
            linear163,
            layernormalization108,
            layernormalization109,
            linear164,
            linear165,
            layernormalization110,
            linear166,
            linear167,
            linear168,
            linear169,
            constant329,
            constant330,
            linear170,
            layernormalization111,
            conv1d55,
            conv1d56,
            layernormalization112,
            conv1d57,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add118_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization107_out1 = {
            let dtype = add118_out1.clone().dtype();
            self.layernormalization107
                .forward(add118_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear162_out1 = self.linear162.forward(layernormalization107_out1);
        let sigmoid72_out1 = burn::tensor::activation::sigmoid(linear162_out1.clone());
        let mul136_out1 = linear162_out1.mul(sigmoid72_out1);
        let linear163_out1 = self.linear163.forward(mul136_out1);
        let mul137_out1 = linear163_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add119_out1 = add118_out1.add(mul137_out1);
        let layernormalization108_out1 = {
            let dtype = add119_out1.dtype();
            self.layernormalization108
                .forward(add119_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather37_out1 = {
            let sliced = cache_last_channel.slice(s![18, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather38_out1 = {
            let sliced = cache_last_time.slice(s![18, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization109_out1 = {
            let dtype = layernormalization108_out1.clone().dtype();
            self.layernormalization109
                .forward(
                    layernormalization108_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear164_out1 = self.linear164.forward(layernormalization109_out1);
        let sigmoid73_out1 = burn::tensor::activation::sigmoid(linear164_out1.clone());
        let mul138_out1 = linear164_out1.mul(sigmoid73_out1);
        let linear165_out1 = self.linear165.forward(mul138_out1);
        let mul139_out1 = linear165_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add120_out1 = layernormalization108_out1.add(mul139_out1);
        let layernormalization110_out1 = {
            let dtype = add120_out1.clone().dtype();
            self.layernormalization110
                .forward(add120_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat55_out1 = burn::tensor::Tensor::cat(
            [gather37_out1.clone(), layernormalization110_out1.clone()].into(),
            1,
        );
        let slice77_out1 = gather37_out1.slice(s![.., 7.., ..]);
        let concat56_out1 = burn::tensor::Tensor::cat(
            [slice77_out1, layernormalization110_out1.clone()].into(),
            1,
        );
        let linear166_out1 = self.linear166.forward(layernormalization110_out1);
        let reshape128_out1 = linear166_out1.reshape([1, -1, 8, 128]);
        let linear167_out1 = self.linear167.forward(concat55_out1.clone());
        let reshape129_out1 = linear167_out1.reshape([1, -1, 8, 128]);
        let linear168_out1 = self.linear168.forward(concat55_out1);
        let reshape130_out1 = linear168_out1.reshape([1, -1, 8, 128]);
        let transpose184_out1 = reshape129_out1.permute([0, 2, 1, 3]);
        let transpose185_out1 = reshape130_out1.permute([0, 2, 1, 3]);
        let linear169_out1 = self.linear169.forward(slice2_out1);
        let reshape131_out1 = linear169_out1.reshape([1, -1, 8, 128]);
        let constant329_out1 = self.constant329.val();
        let add121_out1 = reshape128_out1
            .clone()
            .add((constant329_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose186_out1 = add121_out1.permute([0, 2, 1, 3]);
        let constant330_out1 = self.constant330.val();
        let add122_out1 = reshape128_out1
            .add((constant330_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose187_out1 = add122_out1.permute([0, 2, 1, 3]);
        let transpose188_out1 = reshape131_out1.permute([0, 2, 3, 1]);
        let matmul188_out1 = transpose187_out1.matmul(transpose188_out1);
        let pad37_out1 = matmul188_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape132_out1 = pad37_out1.reshape([1, 8, -1, 7]);
        let slice78_out1 = reshape132_out1.slice(s![.., .., 1.., ..]);
        let reshape133_out1 = slice78_out1.reshape([1, 8, 7, 153]);
        let slice79_out1 = reshape133_out1.slice(s![.., .., .., 0..77]);
        let mul140_out1 = slice79_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where55_out1 = mul140_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention19_out1,) = {
            let q = transpose186_out1;
            let k = transpose184_out1;
            let v = transpose185_out1;
            let score_bias = where55_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention19_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention19_out1 = {
                let dims = attention19_out1.dims();
                attention19_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention19_out1,)
        };
        let where56_out1 = attention19_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose189_out1 = where56_out1.permute([0, 2, 1, 3]);
        let reshape134_out1 = transpose189_out1.reshape([1, -1, 1024]);
        let linear170_out1 = self.linear170.forward(reshape134_out1);
        let add123_out1 = add120_out1.add(linear170_out1);
        let layernormalization111_out1 = {
            let dtype = add123_out1.clone().dtype();
            self.layernormalization111
                .forward(add123_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose190_out1 = layernormalization111_out1.permute([0, 2, 1]);
        let conv1d55_out1 = self.conv1d55.forward(transpose190_out1);
        let split_tensors = conv1d55_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split19_out1, split19_out2] = split_tensors.try_into().unwrap();
        let sigmoid74_out1 = burn::tensor::activation::sigmoid(split19_out2);
        let mul141_out1 = split19_out1.mul(sigmoid74_out1);
        let where57_out1 = mul141_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat57_out1 = burn::tensor::Tensor::cat(
            [gather38_out1, where57_out1].into(),
            2,
        );
        let slice80_out1 = concat57_out1.clone().slice(s![.., .., - 8..]);
        let conv1d56_out1 = self.conv1d56.forward(concat57_out1);
        let transpose191_out1 = conv1d56_out1.permute([0, 2, 1]);
        let layernormalization112_out1 = {
            let dtype = transpose191_out1.dtype();
            self.layernormalization112
                .forward(transpose191_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose192_out1 = layernormalization112_out1.permute([0, 2, 1]);
        let sigmoid75_out1 = burn::tensor::activation::sigmoid(
            transpose192_out1.clone(),
        );
        let mul142_out1 = transpose192_out1.mul(sigmoid75_out1);
        let conv1d57_out1 = self.conv1d57.forward(mul142_out1);
        let transpose193_out1 = conv1d57_out1.permute([0, 2, 1]);
        let add124_out1 = add123_out1.add(transpose193_out1);
        (add124_out1, concat56_out1, slice80_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule22 {
    layernormalization113: LayerNorm,
    linear171: Linear,
    linear172: Linear,
    layernormalization114: LayerNorm,
    layernormalization115: LayerNorm,
    linear173: Linear,
    linear174: Linear,
    layernormalization116: LayerNorm,
    linear175: Linear,
    linear176: Linear,
    linear177: Linear,
    linear178: Linear,
    constant346: burn::module::Param<Tensor<2>>,
    constant347: burn::module::Param<Tensor<2>>,
    linear179: Linear,
    layernormalization117: LayerNorm,
    conv1d58: Conv1d,
    conv1d59: Conv1d,
    layernormalization118: LayerNorm,
    conv1d60: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule22 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization113 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear171 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear172 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization114 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization115 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear173 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear174 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization116 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear175 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear176 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear177 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear178 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant346: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant347: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear179 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization117 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d58 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d59 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization118 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d60 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization113,
            linear171,
            linear172,
            layernormalization114,
            layernormalization115,
            linear173,
            linear174,
            layernormalization116,
            linear175,
            linear176,
            linear177,
            linear178,
            constant346,
            constant347,
            linear179,
            layernormalization117,
            conv1d58,
            conv1d59,
            layernormalization118,
            conv1d60,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add124_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization113_out1 = {
            let dtype = add124_out1.clone().dtype();
            self.layernormalization113
                .forward(add124_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear171_out1 = self.linear171.forward(layernormalization113_out1);
        let sigmoid76_out1 = burn::tensor::activation::sigmoid(linear171_out1.clone());
        let mul143_out1 = linear171_out1.mul(sigmoid76_out1);
        let linear172_out1 = self.linear172.forward(mul143_out1);
        let mul144_out1 = linear172_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add125_out1 = add124_out1.add(mul144_out1);
        let layernormalization114_out1 = {
            let dtype = add125_out1.dtype();
            self.layernormalization114
                .forward(add125_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather39_out1 = {
            let sliced = cache_last_channel.slice(s![19, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather40_out1 = {
            let sliced = cache_last_time.slice(s![19, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization115_out1 = {
            let dtype = layernormalization114_out1.clone().dtype();
            self.layernormalization115
                .forward(
                    layernormalization114_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear173_out1 = self.linear173.forward(layernormalization115_out1);
        let sigmoid77_out1 = burn::tensor::activation::sigmoid(linear173_out1.clone());
        let mul145_out1 = linear173_out1.mul(sigmoid77_out1);
        let linear174_out1 = self.linear174.forward(mul145_out1);
        let mul146_out1 = linear174_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add126_out1 = layernormalization114_out1.add(mul146_out1);
        let layernormalization116_out1 = {
            let dtype = add126_out1.clone().dtype();
            self.layernormalization116
                .forward(add126_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat58_out1 = burn::tensor::Tensor::cat(
            [gather39_out1.clone(), layernormalization116_out1.clone()].into(),
            1,
        );
        let slice81_out1 = gather39_out1.slice(s![.., 7.., ..]);
        let concat59_out1 = burn::tensor::Tensor::cat(
            [slice81_out1, layernormalization116_out1.clone()].into(),
            1,
        );
        let linear175_out1 = self.linear175.forward(layernormalization116_out1);
        let reshape135_out1 = linear175_out1.reshape([1, -1, 8, 128]);
        let linear176_out1 = self.linear176.forward(concat58_out1.clone());
        let reshape136_out1 = linear176_out1.reshape([1, -1, 8, 128]);
        let linear177_out1 = self.linear177.forward(concat58_out1);
        let reshape137_out1 = linear177_out1.reshape([1, -1, 8, 128]);
        let transpose194_out1 = reshape136_out1.permute([0, 2, 1, 3]);
        let transpose195_out1 = reshape137_out1.permute([0, 2, 1, 3]);
        let linear178_out1 = self.linear178.forward(slice2_out1);
        let reshape138_out1 = linear178_out1.reshape([1, -1, 8, 128]);
        let constant346_out1 = self.constant346.val();
        let add127_out1 = reshape135_out1
            .clone()
            .add((constant346_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose196_out1 = add127_out1.permute([0, 2, 1, 3]);
        let constant347_out1 = self.constant347.val();
        let add128_out1 = reshape135_out1
            .add((constant347_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose197_out1 = add128_out1.permute([0, 2, 1, 3]);
        let transpose198_out1 = reshape138_out1.permute([0, 2, 3, 1]);
        let matmul198_out1 = transpose197_out1.matmul(transpose198_out1);
        let pad39_out1 = matmul198_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape139_out1 = pad39_out1.reshape([1, 8, -1, 7]);
        let slice82_out1 = reshape139_out1.slice(s![.., .., 1.., ..]);
        let reshape140_out1 = slice82_out1.reshape([1, 8, 7, 153]);
        let slice83_out1 = reshape140_out1.slice(s![.., .., .., 0..77]);
        let mul147_out1 = slice83_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where58_out1 = mul147_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention20_out1,) = {
            let q = transpose196_out1;
            let k = transpose194_out1;
            let v = transpose195_out1;
            let score_bias = where58_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention20_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention20_out1 = {
                let dims = attention20_out1.dims();
                attention20_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention20_out1,)
        };
        let where59_out1 = attention20_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose199_out1 = where59_out1.permute([0, 2, 1, 3]);
        let reshape141_out1 = transpose199_out1.reshape([1, -1, 1024]);
        let linear179_out1 = self.linear179.forward(reshape141_out1);
        let add129_out1 = add126_out1.add(linear179_out1);
        let layernormalization117_out1 = {
            let dtype = add129_out1.clone().dtype();
            self.layernormalization117
                .forward(add129_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose200_out1 = layernormalization117_out1.permute([0, 2, 1]);
        let conv1d58_out1 = self.conv1d58.forward(transpose200_out1);
        let split_tensors = conv1d58_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split20_out1, split20_out2] = split_tensors.try_into().unwrap();
        let sigmoid78_out1 = burn::tensor::activation::sigmoid(split20_out2);
        let mul148_out1 = split20_out1.mul(sigmoid78_out1);
        let where60_out1 = mul148_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat60_out1 = burn::tensor::Tensor::cat(
            [gather40_out1, where60_out1].into(),
            2,
        );
        let slice84_out1 = concat60_out1.clone().slice(s![.., .., - 8..]);
        let conv1d59_out1 = self.conv1d59.forward(concat60_out1);
        let transpose201_out1 = conv1d59_out1.permute([0, 2, 1]);
        let layernormalization118_out1 = {
            let dtype = transpose201_out1.dtype();
            self.layernormalization118
                .forward(transpose201_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose202_out1 = layernormalization118_out1.permute([0, 2, 1]);
        let sigmoid79_out1 = burn::tensor::activation::sigmoid(
            transpose202_out1.clone(),
        );
        let mul149_out1 = transpose202_out1.mul(sigmoid79_out1);
        let conv1d60_out1 = self.conv1d60.forward(mul149_out1);
        let transpose203_out1 = conv1d60_out1.permute([0, 2, 1]);
        let add130_out1 = add129_out1.add(transpose203_out1);
        (add130_out1, concat59_out1, slice84_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule23 {
    layernormalization119: LayerNorm,
    linear180: Linear,
    linear181: Linear,
    layernormalization120: LayerNorm,
    layernormalization121: LayerNorm,
    linear182: Linear,
    linear183: Linear,
    layernormalization122: LayerNorm,
    linear184: Linear,
    linear185: Linear,
    linear186: Linear,
    linear187: Linear,
    constant363: burn::module::Param<Tensor<2>>,
    constant364: burn::module::Param<Tensor<2>>,
    linear188: Linear,
    layernormalization123: LayerNorm,
    conv1d61: Conv1d,
    conv1d62: Conv1d,
    layernormalization124: LayerNorm,
    conv1d63: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule23 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization119 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear180 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear181 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization120 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization121 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear182 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear183 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization122 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear184 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear185 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear186 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear187 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant363: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant364: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear188 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization123 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d61 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d62 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization124 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d63 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization119,
            linear180,
            linear181,
            layernormalization120,
            layernormalization121,
            linear182,
            linear183,
            layernormalization122,
            linear184,
            linear185,
            linear186,
            linear187,
            constant363,
            constant364,
            linear188,
            layernormalization123,
            conv1d61,
            conv1d62,
            layernormalization124,
            conv1d63,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add130_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization119_out1 = {
            let dtype = add130_out1.clone().dtype();
            self.layernormalization119
                .forward(add130_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear180_out1 = self.linear180.forward(layernormalization119_out1);
        let sigmoid80_out1 = burn::tensor::activation::sigmoid(linear180_out1.clone());
        let mul150_out1 = linear180_out1.mul(sigmoid80_out1);
        let linear181_out1 = self.linear181.forward(mul150_out1);
        let mul151_out1 = linear181_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add131_out1 = add130_out1.add(mul151_out1);
        let layernormalization120_out1 = {
            let dtype = add131_out1.dtype();
            self.layernormalization120
                .forward(add131_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather41_out1 = {
            let sliced = cache_last_channel.slice(s![20, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather42_out1 = {
            let sliced = cache_last_time.slice(s![20, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization121_out1 = {
            let dtype = layernormalization120_out1.clone().dtype();
            self.layernormalization121
                .forward(
                    layernormalization120_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear182_out1 = self.linear182.forward(layernormalization121_out1);
        let sigmoid81_out1 = burn::tensor::activation::sigmoid(linear182_out1.clone());
        let mul152_out1 = linear182_out1.mul(sigmoid81_out1);
        let linear183_out1 = self.linear183.forward(mul152_out1);
        let mul153_out1 = linear183_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add132_out1 = layernormalization120_out1.add(mul153_out1);
        let layernormalization122_out1 = {
            let dtype = add132_out1.clone().dtype();
            self.layernormalization122
                .forward(add132_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat61_out1 = burn::tensor::Tensor::cat(
            [gather41_out1.clone(), layernormalization122_out1.clone()].into(),
            1,
        );
        let slice85_out1 = gather41_out1.slice(s![.., 7.., ..]);
        let concat62_out1 = burn::tensor::Tensor::cat(
            [slice85_out1, layernormalization122_out1.clone()].into(),
            1,
        );
        let linear184_out1 = self.linear184.forward(layernormalization122_out1);
        let reshape142_out1 = linear184_out1.reshape([1, -1, 8, 128]);
        let linear185_out1 = self.linear185.forward(concat61_out1.clone());
        let reshape143_out1 = linear185_out1.reshape([1, -1, 8, 128]);
        let linear186_out1 = self.linear186.forward(concat61_out1);
        let reshape144_out1 = linear186_out1.reshape([1, -1, 8, 128]);
        let transpose204_out1 = reshape143_out1.permute([0, 2, 1, 3]);
        let transpose205_out1 = reshape144_out1.permute([0, 2, 1, 3]);
        let linear187_out1 = self.linear187.forward(slice2_out1);
        let reshape145_out1 = linear187_out1.reshape([1, -1, 8, 128]);
        let constant363_out1 = self.constant363.val();
        let add133_out1 = reshape142_out1
            .clone()
            .add((constant363_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose206_out1 = add133_out1.permute([0, 2, 1, 3]);
        let constant364_out1 = self.constant364.val();
        let add134_out1 = reshape142_out1
            .add((constant364_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose207_out1 = add134_out1.permute([0, 2, 1, 3]);
        let transpose208_out1 = reshape145_out1.permute([0, 2, 3, 1]);
        let matmul208_out1 = transpose207_out1.matmul(transpose208_out1);
        let pad41_out1 = matmul208_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape146_out1 = pad41_out1.reshape([1, 8, -1, 7]);
        let slice86_out1 = reshape146_out1.slice(s![.., .., 1.., ..]);
        let reshape147_out1 = slice86_out1.reshape([1, 8, 7, 153]);
        let slice87_out1 = reshape147_out1.slice(s![.., .., .., 0..77]);
        let mul154_out1 = slice87_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where61_out1 = mul154_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention21_out1,) = {
            let q = transpose206_out1;
            let k = transpose204_out1;
            let v = transpose205_out1;
            let score_bias = where61_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention21_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention21_out1 = {
                let dims = attention21_out1.dims();
                attention21_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention21_out1,)
        };
        let where62_out1 = attention21_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose209_out1 = where62_out1.permute([0, 2, 1, 3]);
        let reshape148_out1 = transpose209_out1.reshape([1, -1, 1024]);
        let linear188_out1 = self.linear188.forward(reshape148_out1);
        let add135_out1 = add132_out1.add(linear188_out1);
        let layernormalization123_out1 = {
            let dtype = add135_out1.clone().dtype();
            self.layernormalization123
                .forward(add135_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose210_out1 = layernormalization123_out1.permute([0, 2, 1]);
        let conv1d61_out1 = self.conv1d61.forward(transpose210_out1);
        let split_tensors = conv1d61_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split21_out1, split21_out2] = split_tensors.try_into().unwrap();
        let sigmoid82_out1 = burn::tensor::activation::sigmoid(split21_out2);
        let mul155_out1 = split21_out1.mul(sigmoid82_out1);
        let where63_out1 = mul155_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat63_out1 = burn::tensor::Tensor::cat(
            [gather42_out1, where63_out1].into(),
            2,
        );
        let slice88_out1 = concat63_out1.clone().slice(s![.., .., - 8..]);
        let conv1d62_out1 = self.conv1d62.forward(concat63_out1);
        let transpose211_out1 = conv1d62_out1.permute([0, 2, 1]);
        let layernormalization124_out1 = {
            let dtype = transpose211_out1.dtype();
            self.layernormalization124
                .forward(transpose211_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose212_out1 = layernormalization124_out1.permute([0, 2, 1]);
        let sigmoid83_out1 = burn::tensor::activation::sigmoid(
            transpose212_out1.clone(),
        );
        let mul156_out1 = transpose212_out1.mul(sigmoid83_out1);
        let conv1d63_out1 = self.conv1d63.forward(mul156_out1);
        let transpose213_out1 = conv1d63_out1.permute([0, 2, 1]);
        let add136_out1 = add135_out1.add(transpose213_out1);
        (add136_out1, concat62_out1, slice88_out1)
    }
}
#[derive(Module, Debug)]
pub struct Submodule24 {
    layernormalization125: LayerNorm,
    linear189: Linear,
    linear190: Linear,
    layernormalization126: LayerNorm,
    layernormalization127: LayerNorm,
    linear191: Linear,
    linear192: Linear,
    layernormalization128: LayerNorm,
    linear193: Linear,
    linear194: Linear,
    linear195: Linear,
    linear196: Linear,
    constant380: burn::module::Param<Tensor<2>>,
    constant381: burn::module::Param<Tensor<2>>,
    linear197: Linear,
    layernormalization129: LayerNorm,
    conv1d64: Conv1d,
    conv1d65: Conv1d,
    layernormalization130: LayerNorm,
    conv1d66: Conv1d,
    layernormalization131: LayerNorm,
    linear198: Linear,
    linear199: Linear,
    layernormalization132: LayerNorm,
    layernormalization133: LayerNorm,
    linear200: Linear,
    linear201: Linear,
    layernormalization134: LayerNorm,
    linear202: Linear,
    linear203: Linear,
    linear204: Linear,
    linear205: Linear,
    constant397: burn::module::Param<Tensor<2>>,
    constant398: burn::module::Param<Tensor<2>>,
    linear206: Linear,
    layernormalization135: LayerNorm,
    conv1d67: Conv1d,
    conv1d68: Conv1d,
    layernormalization136: LayerNorm,
    conv1d69: Conv1d,
    layernormalization137: LayerNorm,
    linear207: Linear,
    linear208: Linear,
    layernormalization138: LayerNorm,
    layernormalization139: LayerNorm,
    linear209: Linear,
    linear210: Linear,
    layernormalization140: LayerNorm,
    linear211: Linear,
    linear212: Linear,
    linear213: Linear,
    linear214: Linear,
    constant414: burn::module::Param<Tensor<2>>,
    constant415: burn::module::Param<Tensor<2>>,
    linear215: Linear,
    layernormalization141: LayerNorm,
    conv1d70: Conv1d,
    #[module(skip)]
    device: Device,
}
impl Submodule24 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let layernormalization125 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear189 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear190 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization126 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization127 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear191 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear192 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization128 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear193 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear194 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear195 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear196 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant380: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant381: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear197 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization129 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d64 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d65 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization130 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d66 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let layernormalization131 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear198 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear199 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization132 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization133 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear200 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear201 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization134 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear202 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear203 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear204 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear205 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant397: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant398: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear206 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization135 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d67 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let conv1d68 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization136 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d69 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let layernormalization137 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear207 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear208 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization138 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let layernormalization139 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear209 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear210 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization140 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear211 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear212 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear213 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let linear214 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let constant414: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let constant415: burn::module::Param<Tensor<2>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                2,
            >::zeros([8, 128], (device, burn::tensor::DType::F32)),
            device.clone(),
            false,
            [8, 128].into(),
        );
        let linear215 = LinearConfig::new(1024, 1024).with_bias(false).init(device);
        let layernormalization141 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d70 = Conv1dConfig::new(1024, 2048, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        Self {
            layernormalization125,
            linear189,
            linear190,
            layernormalization126,
            layernormalization127,
            linear191,
            linear192,
            layernormalization128,
            linear193,
            linear194,
            linear195,
            linear196,
            constant380,
            constant381,
            linear197,
            layernormalization129,
            conv1d64,
            conv1d65,
            layernormalization130,
            conv1d66,
            layernormalization131,
            linear198,
            linear199,
            layernormalization132,
            layernormalization133,
            linear200,
            linear201,
            layernormalization134,
            linear202,
            linear203,
            linear204,
            linear205,
            constant397,
            constant398,
            linear206,
            layernormalization135,
            conv1d67,
            conv1d68,
            layernormalization136,
            conv1d69,
            layernormalization137,
            linear207,
            linear208,
            layernormalization138,
            layernormalization139,
            linear209,
            linear210,
            layernormalization140,
            linear211,
            linear212,
            linear213,
            linear214,
            constant414,
            constant415,
            linear215,
            layernormalization141,
            conv1d70,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        add136_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        slice2_out1: Tensor<3>,
        constant672_out1: Tensor<1>,
        unsqueeze20_out1: Tensor<4, Bool>,
        constant444_out1: f32,
        expand12_out1: Tensor<4, Bool>,
        constant447_out1: f32,
        unsqueeze22_out1: Tensor<3, Bool>,
    ) -> (Tensor<3>, Tensor<3>, Tensor<3>, Tensor<3>, Tensor<3>, Tensor<3>, Tensor<3>) {
        let layernormalization125_out1 = {
            let dtype = add136_out1.clone().dtype();
            self.layernormalization125
                .forward(add136_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear189_out1 = self.linear189.forward(layernormalization125_out1);
        let sigmoid84_out1 = burn::tensor::activation::sigmoid(linear189_out1.clone());
        let mul157_out1 = linear189_out1.mul(sigmoid84_out1);
        let linear190_out1 = self.linear190.forward(mul157_out1);
        let mul158_out1 = linear190_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add137_out1 = add136_out1.add(mul158_out1);
        let layernormalization126_out1 = {
            let dtype = add137_out1.dtype();
            self.layernormalization126
                .forward(add137_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather43_out1 = {
            let sliced = cache_last_channel.clone().slice(s![21, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather44_out1 = {
            let sliced = cache_last_time.clone().slice(s![21, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization127_out1 = {
            let dtype = layernormalization126_out1.clone().dtype();
            self.layernormalization127
                .forward(
                    layernormalization126_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear191_out1 = self.linear191.forward(layernormalization127_out1);
        let sigmoid85_out1 = burn::tensor::activation::sigmoid(linear191_out1.clone());
        let mul159_out1 = linear191_out1.mul(sigmoid85_out1);
        let linear192_out1 = self.linear192.forward(mul159_out1);
        let mul160_out1 = linear192_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add138_out1 = layernormalization126_out1.add(mul160_out1);
        let layernormalization128_out1 = {
            let dtype = add138_out1.clone().dtype();
            self.layernormalization128
                .forward(add138_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat64_out1 = burn::tensor::Tensor::cat(
            [gather43_out1.clone(), layernormalization128_out1.clone()].into(),
            1,
        );
        let slice89_out1 = gather43_out1.slice(s![.., 7.., ..]);
        let concat65_out1 = burn::tensor::Tensor::cat(
            [slice89_out1, layernormalization128_out1.clone()].into(),
            1,
        );
        let linear193_out1 = self.linear193.forward(layernormalization128_out1);
        let reshape149_out1 = linear193_out1.reshape([1, -1, 8, 128]);
        let linear194_out1 = self.linear194.forward(concat64_out1.clone());
        let reshape150_out1 = linear194_out1.reshape([1, -1, 8, 128]);
        let linear195_out1 = self.linear195.forward(concat64_out1);
        let reshape151_out1 = linear195_out1.reshape([1, -1, 8, 128]);
        let transpose214_out1 = reshape150_out1.permute([0, 2, 1, 3]);
        let transpose215_out1 = reshape151_out1.permute([0, 2, 1, 3]);
        let linear196_out1 = self.linear196.forward(slice2_out1.clone());
        let reshape152_out1 = linear196_out1.reshape([1, -1, 8, 128]);
        let constant380_out1 = self.constant380.val();
        let add139_out1 = reshape149_out1
            .clone()
            .add((constant380_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose216_out1 = add139_out1.permute([0, 2, 1, 3]);
        let constant381_out1 = self.constant381.val();
        let add140_out1 = reshape149_out1
            .add((constant381_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose217_out1 = add140_out1.permute([0, 2, 1, 3]);
        let transpose218_out1 = reshape152_out1.permute([0, 2, 3, 1]);
        let matmul218_out1 = transpose217_out1.matmul(transpose218_out1);
        let pad43_out1 = matmul218_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape153_out1 = pad43_out1.reshape([1, 8, -1, 7]);
        let slice90_out1 = reshape153_out1.slice(s![.., .., 1.., ..]);
        let reshape154_out1 = slice90_out1.reshape([1, 8, 7, 153]);
        let slice91_out1 = reshape154_out1.slice(s![.., .., .., 0..77]);
        let mul161_out1 = slice91_out1
            .mul((constant672_out1.clone()).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where64_out1 = mul161_out1
            .mask_fill(unsqueeze20_out1.clone(), constant444_out1);
        let (attention22_out1,) = {
            let q = transpose216_out1;
            let k = transpose214_out1;
            let v = transpose215_out1;
            let score_bias = where64_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention22_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention22_out1 = {
                let dims = attention22_out1.dims();
                attention22_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention22_out1,)
        };
        let where65_out1 = attention22_out1
            .mask_fill(expand12_out1.clone(), constant447_out1);
        let transpose219_out1 = where65_out1.permute([0, 2, 1, 3]);
        let reshape155_out1 = transpose219_out1.reshape([1, -1, 1024]);
        let linear197_out1 = self.linear197.forward(reshape155_out1);
        let add141_out1 = add138_out1.add(linear197_out1);
        let layernormalization129_out1 = {
            let dtype = add141_out1.clone().dtype();
            self.layernormalization129
                .forward(add141_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose220_out1 = layernormalization129_out1.permute([0, 2, 1]);
        let conv1d64_out1 = self.conv1d64.forward(transpose220_out1);
        let split_tensors = conv1d64_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split22_out1, split22_out2] = split_tensors.try_into().unwrap();
        let sigmoid86_out1 = burn::tensor::activation::sigmoid(split22_out2);
        let mul162_out1 = split22_out1.mul(sigmoid86_out1);
        let where66_out1 = mul162_out1
            .mask_fill(unsqueeze22_out1.clone(), constant447_out1);
        let concat66_out1 = burn::tensor::Tensor::cat(
            [gather44_out1, where66_out1].into(),
            2,
        );
        let slice92_out1 = concat66_out1.clone().slice(s![.., .., - 8..]);
        let conv1d65_out1 = self.conv1d65.forward(concat66_out1);
        let transpose221_out1 = conv1d65_out1.permute([0, 2, 1]);
        let layernormalization130_out1 = {
            let dtype = transpose221_out1.dtype();
            self.layernormalization130
                .forward(transpose221_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose222_out1 = layernormalization130_out1.permute([0, 2, 1]);
        let sigmoid87_out1 = burn::tensor::activation::sigmoid(
            transpose222_out1.clone(),
        );
        let mul163_out1 = transpose222_out1.mul(sigmoid87_out1);
        let conv1d66_out1 = self.conv1d66.forward(mul163_out1);
        let transpose223_out1 = conv1d66_out1.permute([0, 2, 1]);
        let add142_out1 = add141_out1.add(transpose223_out1);
        let layernormalization131_out1 = {
            let dtype = add142_out1.clone().dtype();
            self.layernormalization131
                .forward(add142_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear198_out1 = self.linear198.forward(layernormalization131_out1);
        let sigmoid88_out1 = burn::tensor::activation::sigmoid(linear198_out1.clone());
        let mul164_out1 = linear198_out1.mul(sigmoid88_out1);
        let linear199_out1 = self.linear199.forward(mul164_out1);
        let mul165_out1 = linear199_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add143_out1 = add142_out1.add(mul165_out1);
        let layernormalization132_out1 = {
            let dtype = add143_out1.dtype();
            self.layernormalization132
                .forward(add143_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather45_out1 = {
            let sliced = cache_last_channel.clone().slice(s![22, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather46_out1 = {
            let sliced = cache_last_time.clone().slice(s![22, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization133_out1 = {
            let dtype = layernormalization132_out1.clone().dtype();
            self.layernormalization133
                .forward(
                    layernormalization132_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear200_out1 = self.linear200.forward(layernormalization133_out1);
        let sigmoid89_out1 = burn::tensor::activation::sigmoid(linear200_out1.clone());
        let mul166_out1 = linear200_out1.mul(sigmoid89_out1);
        let linear201_out1 = self.linear201.forward(mul166_out1);
        let mul167_out1 = linear201_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add144_out1 = layernormalization132_out1.add(mul167_out1);
        let layernormalization134_out1 = {
            let dtype = add144_out1.clone().dtype();
            self.layernormalization134
                .forward(add144_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat67_out1 = burn::tensor::Tensor::cat(
            [gather45_out1.clone(), layernormalization134_out1.clone()].into(),
            1,
        );
        let slice93_out1 = gather45_out1.slice(s![.., 7.., ..]);
        let concat68_out1 = burn::tensor::Tensor::cat(
            [slice93_out1, layernormalization134_out1.clone()].into(),
            1,
        );
        let linear202_out1 = self.linear202.forward(layernormalization134_out1);
        let reshape156_out1 = linear202_out1.reshape([1, -1, 8, 128]);
        let linear203_out1 = self.linear203.forward(concat67_out1.clone());
        let reshape157_out1 = linear203_out1.reshape([1, -1, 8, 128]);
        let linear204_out1 = self.linear204.forward(concat67_out1);
        let reshape158_out1 = linear204_out1.reshape([1, -1, 8, 128]);
        let transpose224_out1 = reshape157_out1.permute([0, 2, 1, 3]);
        let transpose225_out1 = reshape158_out1.permute([0, 2, 1, 3]);
        let linear205_out1 = self.linear205.forward(slice2_out1.clone());
        let reshape159_out1 = linear205_out1.reshape([1, -1, 8, 128]);
        let constant397_out1 = self.constant397.val();
        let add145_out1 = reshape156_out1
            .clone()
            .add((constant397_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose226_out1 = add145_out1.permute([0, 2, 1, 3]);
        let constant398_out1 = self.constant398.val();
        let add146_out1 = reshape156_out1
            .add((constant398_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose227_out1 = add146_out1.permute([0, 2, 1, 3]);
        let transpose228_out1 = reshape159_out1.permute([0, 2, 3, 1]);
        let matmul228_out1 = transpose227_out1.matmul(transpose228_out1);
        let pad45_out1 = matmul228_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape160_out1 = pad45_out1.reshape([1, 8, -1, 7]);
        let slice94_out1 = reshape160_out1.slice(s![.., .., 1.., ..]);
        let reshape161_out1 = slice94_out1.reshape([1, 8, 7, 153]);
        let slice95_out1 = reshape161_out1.slice(s![.., .., .., 0..77]);
        let mul168_out1 = slice95_out1
            .mul((constant672_out1.clone()).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where67_out1 = mul168_out1
            .mask_fill(unsqueeze20_out1.clone(), constant444_out1);
        let (attention23_out1,) = {
            let q = transpose226_out1;
            let k = transpose224_out1;
            let v = transpose225_out1;
            let score_bias = where67_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention23_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention23_out1 = {
                let dims = attention23_out1.dims();
                attention23_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention23_out1,)
        };
        let where68_out1 = attention23_out1
            .mask_fill(expand12_out1.clone(), constant447_out1);
        let transpose229_out1 = where68_out1.permute([0, 2, 1, 3]);
        let reshape162_out1 = transpose229_out1.reshape([1, -1, 1024]);
        let linear206_out1 = self.linear206.forward(reshape162_out1);
        let add147_out1 = add144_out1.add(linear206_out1);
        let layernormalization135_out1 = {
            let dtype = add147_out1.clone().dtype();
            self.layernormalization135
                .forward(add147_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose230_out1 = layernormalization135_out1.permute([0, 2, 1]);
        let conv1d67_out1 = self.conv1d67.forward(transpose230_out1);
        let split_tensors = conv1d67_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split23_out1, split23_out2] = split_tensors.try_into().unwrap();
        let sigmoid90_out1 = burn::tensor::activation::sigmoid(split23_out2);
        let mul169_out1 = split23_out1.mul(sigmoid90_out1);
        let where69_out1 = mul169_out1
            .mask_fill(unsqueeze22_out1.clone(), constant447_out1);
        let concat69_out1 = burn::tensor::Tensor::cat(
            [gather46_out1, where69_out1].into(),
            2,
        );
        let slice96_out1 = concat69_out1.clone().slice(s![.., .., - 8..]);
        let conv1d68_out1 = self.conv1d68.forward(concat69_out1);
        let transpose231_out1 = conv1d68_out1.permute([0, 2, 1]);
        let layernormalization136_out1 = {
            let dtype = transpose231_out1.dtype();
            self.layernormalization136
                .forward(transpose231_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose232_out1 = layernormalization136_out1.permute([0, 2, 1]);
        let sigmoid91_out1 = burn::tensor::activation::sigmoid(
            transpose232_out1.clone(),
        );
        let mul170_out1 = transpose232_out1.mul(sigmoid91_out1);
        let conv1d69_out1 = self.conv1d69.forward(mul170_out1);
        let transpose233_out1 = conv1d69_out1.permute([0, 2, 1]);
        let add148_out1 = add147_out1.add(transpose233_out1);
        let layernormalization137_out1 = {
            let dtype = add148_out1.clone().dtype();
            self.layernormalization137
                .forward(add148_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear207_out1 = self.linear207.forward(layernormalization137_out1);
        let sigmoid92_out1 = burn::tensor::activation::sigmoid(linear207_out1.clone());
        let mul171_out1 = linear207_out1.mul(sigmoid92_out1);
        let linear208_out1 = self.linear208.forward(mul171_out1);
        let mul172_out1 = linear208_out1
            .mul((constant670_out1.clone()).unsqueeze_dims(&[0isize, 1isize]));
        let add149_out1 = add148_out1.add(mul172_out1);
        let layernormalization138_out1 = {
            let dtype = add149_out1.dtype();
            self.layernormalization138
                .forward(add149_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let gather47_out1 = {
            let sliced = cache_last_channel.slice(s![23, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let gather48_out1 = {
            let sliced = cache_last_time.slice(s![23, .., .., ..]);
            sliced.squeeze_dim::<3usize>(0)
        };
        let layernormalization139_out1 = {
            let dtype = layernormalization138_out1.clone().dtype();
            self.layernormalization139
                .forward(
                    layernormalization138_out1.clone().cast(burn::tensor::DType::F32),
                )
                .cast(dtype)
        };
        let linear209_out1 = self.linear209.forward(layernormalization139_out1);
        let sigmoid93_out1 = burn::tensor::activation::sigmoid(linear209_out1.clone());
        let mul173_out1 = linear209_out1.mul(sigmoid93_out1);
        let linear210_out1 = self.linear210.forward(mul173_out1);
        let mul174_out1 = linear210_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add150_out1 = layernormalization138_out1.add(mul174_out1);
        let layernormalization140_out1 = {
            let dtype = add150_out1.clone().dtype();
            self.layernormalization140
                .forward(add150_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let concat70_out1 = burn::tensor::Tensor::cat(
            [gather47_out1.clone(), layernormalization140_out1.clone()].into(),
            1,
        );
        let slice97_out1 = gather47_out1.slice(s![.., 7.., ..]);
        let concat71_out1 = burn::tensor::Tensor::cat(
            [slice97_out1, layernormalization140_out1.clone()].into(),
            1,
        );
        let linear211_out1 = self.linear211.forward(layernormalization140_out1);
        let reshape163_out1 = linear211_out1.reshape([1, -1, 8, 128]);
        let linear212_out1 = self.linear212.forward(concat70_out1.clone());
        let reshape164_out1 = linear212_out1.reshape([1, -1, 8, 128]);
        let linear213_out1 = self.linear213.forward(concat70_out1);
        let reshape165_out1 = linear213_out1.reshape([1, -1, 8, 128]);
        let transpose234_out1 = reshape164_out1.permute([0, 2, 1, 3]);
        let transpose235_out1 = reshape165_out1.permute([0, 2, 1, 3]);
        let linear214_out1 = self.linear214.forward(slice2_out1);
        let reshape166_out1 = linear214_out1.reshape([1, -1, 8, 128]);
        let constant414_out1 = self.constant414.val();
        let add151_out1 = reshape163_out1
            .clone()
            .add((constant414_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose236_out1 = add151_out1.permute([0, 2, 1, 3]);
        let constant415_out1 = self.constant415.val();
        let add152_out1 = reshape163_out1
            .add((constant415_out1).unsqueeze_dims(&[0isize, 1isize]));
        let transpose237_out1 = add152_out1.permute([0, 2, 1, 3]);
        let transpose238_out1 = reshape166_out1.permute([0, 2, 3, 1]);
        let matmul238_out1 = transpose237_out1.matmul(transpose238_out1);
        let pad47_out1 = matmul238_out1
            .pad(
                [(0usize, 0usize), (0usize, 0usize), (0usize, 0usize), (1usize, 0usize)],
                burn::tensor::ops::PadMode::Constant(0f32),
            );
        let reshape167_out1 = pad47_out1.reshape([1, 8, -1, 7]);
        let slice98_out1 = reshape167_out1.slice(s![.., .., 1.., ..]);
        let reshape168_out1 = slice98_out1.reshape([1, 8, 7, 153]);
        let slice99_out1 = reshape168_out1.slice(s![.., .., .., 0..77]);
        let mul175_out1 = slice99_out1
            .mul((constant672_out1).unsqueeze_dims(&[0isize, 1isize, 2isize]));
        let where70_out1 = mul175_out1.mask_fill(unsqueeze20_out1, constant444_out1);
        let (attention24_out1,) = {
            let q = transpose236_out1;
            let k = transpose234_out1;
            let v = transpose235_out1;
            let score_bias = where70_out1;
            let hidden_rows = score_bias
                .clone()
                .max_dim(3)
                .equal_elem(f32::NEG_INFINITY);
            let attention24_out1 = burn::tensor::module::attention(
                q,
                k,
                v,
                None,
                Some(score_bias),
                burn::tensor::ops::AttentionModuleOptions {
                    scale: None,
                    softcap: None,
                    is_causal: false,
                },
            );
            let attention24_out1 = {
                let dims = attention24_out1.dims();
                attention24_out1.mask_fill(hidden_rows.expand(dims), 0.0)
            };
            (attention24_out1,)
        };
        let where71_out1 = attention24_out1.mask_fill(expand12_out1, constant447_out1);
        let transpose239_out1 = where71_out1.permute([0, 2, 1, 3]);
        let reshape169_out1 = transpose239_out1.reshape([1, -1, 1024]);
        let linear215_out1 = self.linear215.forward(reshape169_out1);
        let add153_out1 = add150_out1.add(linear215_out1);
        let layernormalization141_out1 = {
            let dtype = add153_out1.clone().dtype();
            self.layernormalization141
                .forward(add153_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose240_out1 = layernormalization141_out1.permute([0, 2, 1]);
        let conv1d70_out1 = self.conv1d70.forward(transpose240_out1);
        let split_tensors = conv1d70_out1.split_with_sizes([1024, 1024].into(), 1);
        let [split24_out1, split24_out2] = split_tensors.try_into().unwrap();
        let sigmoid94_out1 = burn::tensor::activation::sigmoid(split24_out2);
        let mul176_out1 = split24_out1.mul(sigmoid94_out1);
        let where72_out1 = mul176_out1.mask_fill(unsqueeze22_out1, constant447_out1);
        let concat72_out1 = burn::tensor::Tensor::cat(
            [gather48_out1, where72_out1].into(),
            2,
        );
        (
            concat72_out1,
            add153_out1,
            concat65_out1,
            concat68_out1,
            concat71_out1,
            slice92_out1,
            slice96_out1,
        )
    }
}
#[derive(Module, Debug)]
pub struct Submodule25 {
    conv1d71: Conv1d,
    layernormalization142: LayerNorm,
    conv1d72: Conv1d,
    layernormalization143: LayerNorm,
    linear216: Linear,
    linear217: Linear,
    layernormalization144: LayerNorm,
    constant692: burn::module::Param<Tensor<1, Int>>,
    #[module(skip)]
    device: Device,
}
impl Submodule25 {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let conv1d71 = Conv1dConfig::new(1024, 1024, 9)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1024)
            .with_bias(false)
            .init(device);
        let layernormalization142 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let conv1d72 = Conv1dConfig::new(1024, 1024, 1)
            .with_stride(1)
            .with_padding(PaddingConfig1d::Valid)
            .with_dilation(1)
            .with_groups(1)
            .with_bias(false)
            .init(device);
        let layernormalization143 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let linear216 = LinearConfig::new(1024, 4096).with_bias(false).init(device);
        let linear217 = LinearConfig::new(4096, 1024).with_bias(false).init(device);
        let layernormalization144 = LayerNormConfig::new(1024)
            .with_epsilon(0.000009999999747378752f64)
            .with_bias(true)
            .init(device);
        let constant692: burn::module::Param<Tensor<1, Int>> = burn::module::Param::uninitialized(
            burn::module::ParamId::new(),
            move |device, _require_grad| Tensor::<
                1,
                Int,
            >::from_data(
                burn::tensor::TensorData::from([7i64]),
                (device, burn::tensor::DType::I64),
            ),
            device.clone(),
            false,
            [1].into(),
        );
        Self {
            conv1d71,
            layernormalization142,
            conv1d72,
            layernormalization143,
            linear216,
            linear217,
            layernormalization144,
            constant692,
            device: device.clone(),
        }
    }
    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        concat72_out1: Tensor<3>,
        add153_out1: Tensor<3>,
        constant670_out1: Tensor<1>,
        concat2_out1: Tensor<3>,
        concat5_out1: Tensor<3>,
        concat8_out1: Tensor<3>,
        concat11_out1: Tensor<3>,
        concat14_out1: Tensor<3>,
        concat17_out1: Tensor<3>,
        concat20_out1: Tensor<3>,
        concat23_out1: Tensor<3>,
        concat26_out1: Tensor<3>,
        concat29_out1: Tensor<3>,
        concat32_out1: Tensor<3>,
        concat35_out1: Tensor<3>,
        concat38_out1: Tensor<3>,
        concat41_out1: Tensor<3>,
        concat44_out1: Tensor<3>,
        concat47_out1: Tensor<3>,
        concat50_out1: Tensor<3>,
        concat53_out1: Tensor<3>,
        concat56_out1: Tensor<3>,
        concat59_out1: Tensor<3>,
        concat62_out1: Tensor<3>,
        concat65_out1: Tensor<3>,
        concat68_out1: Tensor<3>,
        concat71_out1: Tensor<3>,
        slice8_out1: Tensor<3>,
        slice12_out1: Tensor<3>,
        slice16_out1: Tensor<3>,
        slice20_out1: Tensor<3>,
        slice24_out1: Tensor<3>,
        slice28_out1: Tensor<3>,
        slice32_out1: Tensor<3>,
        slice36_out1: Tensor<3>,
        slice40_out1: Tensor<3>,
        slice44_out1: Tensor<3>,
        slice48_out1: Tensor<3>,
        slice52_out1: Tensor<3>,
        slice56_out1: Tensor<3>,
        slice60_out1: Tensor<3>,
        slice64_out1: Tensor<3>,
        slice68_out1: Tensor<3>,
        slice72_out1: Tensor<3>,
        slice76_out1: Tensor<3>,
        slice80_out1: Tensor<3>,
        slice84_out1: Tensor<3>,
        slice88_out1: Tensor<3>,
        slice92_out1: Tensor<3>,
        slice96_out1: Tensor<3>,
        cache_last_channel_len: Tensor<1, Int>,
        clip1_out1: Tensor<1, Int>,
    ) -> (Tensor<3>, Tensor<1, Int>, Tensor<4>, Tensor<4>, Tensor<1, Int>) {
        let slice100_out1 = concat72_out1.clone().slice(s![.., .., - 8..]);
        let conv1d71_out1 = self.conv1d71.forward(concat72_out1);
        let transpose241_out1 = conv1d71_out1.permute([0, 2, 1]);
        let layernormalization142_out1 = {
            let dtype = transpose241_out1.dtype();
            self.layernormalization142
                .forward(transpose241_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose242_out1 = layernormalization142_out1.permute([0, 2, 1]);
        let sigmoid95_out1 = burn::tensor::activation::sigmoid(
            transpose242_out1.clone(),
        );
        let mul177_out1 = transpose242_out1.mul(sigmoid95_out1);
        let conv1d72_out1 = self.conv1d72.forward(mul177_out1);
        let transpose243_out1 = conv1d72_out1.permute([0, 2, 1]);
        let add154_out1 = add153_out1.add(transpose243_out1);
        let layernormalization143_out1 = {
            let dtype = add154_out1.clone().dtype();
            self.layernormalization143
                .forward(add154_out1.clone().cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let linear216_out1 = self.linear216.forward(layernormalization143_out1);
        let sigmoid96_out1 = burn::tensor::activation::sigmoid(linear216_out1.clone());
        let mul178_out1 = linear216_out1.mul(sigmoid96_out1);
        let linear217_out1 = self.linear217.forward(mul178_out1);
        let mul179_out1 = linear217_out1
            .mul((constant670_out1).unsqueeze_dims(&[0isize, 1isize]));
        let add155_out1 = add154_out1.add(mul179_out1);
        let layernormalization144_out1 = {
            let dtype = add155_out1.dtype();
            self.layernormalization144
                .forward(add155_out1.cast(burn::tensor::DType::F32))
                .cast(dtype)
        };
        let transpose244_out1 = layernormalization144_out1.permute([0, 2, 1]);
        let unsqueeze23_out1: Tensor<4> = concat2_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze24_out1: Tensor<4> = concat5_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze25_out1: Tensor<4> = concat8_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze26_out1: Tensor<4> = concat11_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze27_out1: Tensor<4> = concat14_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze28_out1: Tensor<4> = concat17_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze29_out1: Tensor<4> = concat20_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze30_out1: Tensor<4> = concat23_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze31_out1: Tensor<4> = concat26_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze32_out1: Tensor<4> = concat29_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze33_out1: Tensor<4> = concat32_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze34_out1: Tensor<4> = concat35_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze35_out1: Tensor<4> = concat38_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze36_out1: Tensor<4> = concat41_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze37_out1: Tensor<4> = concat44_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze38_out1: Tensor<4> = concat47_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze39_out1: Tensor<4> = concat50_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze40_out1: Tensor<4> = concat53_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze41_out1: Tensor<4> = concat56_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze42_out1: Tensor<4> = concat59_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze43_out1: Tensor<4> = concat62_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze44_out1: Tensor<4> = concat65_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze45_out1: Tensor<4> = concat68_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze46_out1: Tensor<4> = concat71_out1.unsqueeze_dims::<4>(&[0]);
        let concat73_out1 = burn::tensor::Tensor::cat(
            [
                unsqueeze23_out1,
                unsqueeze24_out1,
                unsqueeze25_out1,
                unsqueeze26_out1,
                unsqueeze27_out1,
                unsqueeze28_out1,
                unsqueeze29_out1,
                unsqueeze30_out1,
                unsqueeze31_out1,
                unsqueeze32_out1,
                unsqueeze33_out1,
                unsqueeze34_out1,
                unsqueeze35_out1,
                unsqueeze36_out1,
                unsqueeze37_out1,
                unsqueeze38_out1,
                unsqueeze39_out1,
                unsqueeze40_out1,
                unsqueeze41_out1,
                unsqueeze42_out1,
                unsqueeze43_out1,
                unsqueeze44_out1,
                unsqueeze45_out1,
                unsqueeze46_out1,
            ]
                .into(),
            0,
        );
        let unsqueeze47_out1: Tensor<4> = slice8_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze48_out1: Tensor<4> = slice12_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze49_out1: Tensor<4> = slice16_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze50_out1: Tensor<4> = slice20_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze51_out1: Tensor<4> = slice24_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze52_out1: Tensor<4> = slice28_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze53_out1: Tensor<4> = slice32_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze54_out1: Tensor<4> = slice36_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze55_out1: Tensor<4> = slice40_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze56_out1: Tensor<4> = slice44_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze57_out1: Tensor<4> = slice48_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze58_out1: Tensor<4> = slice52_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze59_out1: Tensor<4> = slice56_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze60_out1: Tensor<4> = slice60_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze61_out1: Tensor<4> = slice64_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze62_out1: Tensor<4> = slice68_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze63_out1: Tensor<4> = slice72_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze64_out1: Tensor<4> = slice76_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze65_out1: Tensor<4> = slice80_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze66_out1: Tensor<4> = slice84_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze67_out1: Tensor<4> = slice88_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze68_out1: Tensor<4> = slice92_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze69_out1: Tensor<4> = slice96_out1.unsqueeze_dims::<4>(&[0]);
        let unsqueeze70_out1: Tensor<4> = slice100_out1.unsqueeze_dims::<4>(&[0]);
        let concat74_out1 = burn::tensor::Tensor::cat(
            [
                unsqueeze47_out1,
                unsqueeze48_out1,
                unsqueeze49_out1,
                unsqueeze50_out1,
                unsqueeze51_out1,
                unsqueeze52_out1,
                unsqueeze53_out1,
                unsqueeze54_out1,
                unsqueeze55_out1,
                unsqueeze56_out1,
                unsqueeze57_out1,
                unsqueeze58_out1,
                unsqueeze59_out1,
                unsqueeze60_out1,
                unsqueeze61_out1,
                unsqueeze62_out1,
                unsqueeze63_out1,
                unsqueeze64_out1,
                unsqueeze65_out1,
                unsqueeze66_out1,
                unsqueeze67_out1,
                unsqueeze68_out1,
                unsqueeze69_out1,
                unsqueeze70_out1,
            ]
                .into(),
            0,
        );
        let constant692_out1 = self.constant692.val();
        let add156_out1 = cache_last_channel_len.add(constant692_out1);
        let clip2_out1 = add156_out1.clamp_max(70f64);
        let slice101_out1 = concat73_out1.slice(s![.., .., - 70.., ..]);
        let slice102_out1 = transpose244_out1.slice(s![.., .., 0..7]);
        let clip3_out1 = clip1_out1.clamp_max(7f64);
        (slice102_out1, clip3_out1, slice101_out1, concat74_out1, clip2_out1)
    }
}

#[derive(Module, Debug)]
pub struct Model {
    submodule1: Submodule1,
    submodule2: Submodule2,
    submodule3: Submodule3,
    submodule4: Submodule4,
    submodule5: Submodule5,
    submodule6: Submodule6,
    submodule7: Submodule7,
    submodule8: Submodule8,
    submodule9: Submodule9,
    submodule10: Submodule10,
    submodule11: Submodule11,
    submodule12: Submodule12,
    submodule13: Submodule13,
    submodule14: Submodule14,
    submodule15: Submodule15,
    submodule16: Submodule16,
    submodule17: Submodule17,
    submodule18: Submodule18,
    submodule19: Submodule19,
    submodule20: Submodule20,
    submodule21: Submodule21,
    submodule22: Submodule22,
    submodule23: Submodule23,
    submodule24: Submodule24,
    submodule25: Submodule25,
    #[module(skip)]
    device: Device,
}





impl Model {
    #[allow(unused_variables)]
    pub fn new(device: &Device) -> Self {
        let submodule1 = Submodule1::new(device);
        let submodule2 = Submodule2::new(device);
        let submodule3 = Submodule3::new(device);
        let submodule4 = Submodule4::new(device);
        let submodule5 = Submodule5::new(device);
        let submodule6 = Submodule6::new(device);
        let submodule7 = Submodule7::new(device);
        let submodule8 = Submodule8::new(device);
        let submodule9 = Submodule9::new(device);
        let submodule10 = Submodule10::new(device);
        let submodule11 = Submodule11::new(device);
        let submodule12 = Submodule12::new(device);
        let submodule13 = Submodule13::new(device);
        let submodule14 = Submodule14::new(device);
        let submodule15 = Submodule15::new(device);
        let submodule16 = Submodule16::new(device);
        let submodule17 = Submodule17::new(device);
        let submodule18 = Submodule18::new(device);
        let submodule19 = Submodule19::new(device);
        let submodule20 = Submodule20::new(device);
        let submodule21 = Submodule21::new(device);
        let submodule22 = Submodule22::new(device);
        let submodule23 = Submodule23::new(device);
        let submodule24 = Submodule24::new(device);
        let submodule25 = Submodule25::new(device);
        Self {
            submodule1,
            submodule2,
            submodule3,
            submodule4,
            submodule5,
            submodule6,
            submodule7,
            submodule8,
            submodule9,
            submodule10,
            submodule11,
            submodule12,
            submodule13,
            submodule14,
            submodule15,
            submodule16,
            submodule17,
            submodule18,
            submodule19,
            submodule20,
            submodule21,
            submodule22,
            submodule23,
            submodule24,
            submodule25,
            device: device.clone(),
        }
    }

    #[allow(clippy::let_and_return, clippy::approx_constant)]
    pub fn forward(
        &self,
        processed_signal: Tensor<3>,
        processed_signal_length: Tensor<1, Int>,
        cache_last_channel: Tensor<4>,
        cache_last_time: Tensor<4>,
        cache_last_channel_len: Tensor<1, Int>,
    ) -> (Tensor<3>, Tensor<1, Int>, Tensor<4>, Tensor<4>, Tensor<1, Int>) {
        let (cast8_out1, conv2d4_out1) = self
            .submodule1
            .forward(processed_signal, processed_signal_length);
        let (slice1_out1, slice2_out1, slice4_out1, slice3_out1, clip1_out1) = self
            .submodule2
            .forward(cast8_out1, conv2d4_out1, cache_last_channel_len.clone());
        let (
            add16_out1,
            constant670_out1,
            constant672_out1,
            unsqueeze20_out1,
            constant444_out1,
            expand12_out1,
            constant447_out1,
            unsqueeze22_out1,
            concat2_out1,
            slice8_out1,
        ) = self
            .submodule3
            .forward(
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice1_out1,
                slice2_out1.clone(),
                slice4_out1,
                slice3_out1,
            );
        let (add22_out1, concat5_out1, slice12_out1) = self
            .submodule4
            .forward(
                add16_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add28_out1, concat8_out1, slice16_out1) = self
            .submodule5
            .forward(
                add22_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add34_out1, concat11_out1, slice20_out1) = self
            .submodule6
            .forward(
                add28_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add40_out1, concat14_out1, slice24_out1) = self
            .submodule7
            .forward(
                add34_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add46_out1, concat17_out1, slice28_out1) = self
            .submodule8
            .forward(
                add40_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add52_out1, concat20_out1, slice32_out1) = self
            .submodule9
            .forward(
                add46_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add58_out1, concat23_out1, slice36_out1) = self
            .submodule10
            .forward(
                add52_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add64_out1, concat26_out1, slice40_out1) = self
            .submodule11
            .forward(
                add58_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add70_out1, concat29_out1, slice44_out1) = self
            .submodule12
            .forward(
                add64_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add76_out1, concat32_out1, slice48_out1) = self
            .submodule13
            .forward(
                add70_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add82_out1, concat35_out1, slice52_out1) = self
            .submodule14
            .forward(
                add76_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add88_out1, concat38_out1, slice56_out1) = self
            .submodule15
            .forward(
                add82_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add94_out1, concat41_out1, slice60_out1) = self
            .submodule16
            .forward(
                add88_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add100_out1, concat44_out1, slice64_out1) = self
            .submodule17
            .forward(
                add94_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add106_out1, concat47_out1, slice68_out1) = self
            .submodule18
            .forward(
                add100_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add112_out1, concat50_out1, slice72_out1) = self
            .submodule19
            .forward(
                add106_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add118_out1, concat53_out1, slice76_out1) = self
            .submodule20
            .forward(
                add112_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add124_out1, concat56_out1, slice80_out1) = self
            .submodule21
            .forward(
                add118_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add130_out1, concat59_out1, slice84_out1) = self
            .submodule22
            .forward(
                add124_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (add136_out1, concat62_out1, slice88_out1) = self
            .submodule23
            .forward(
                add130_out1,
                constant670_out1.clone(),
                cache_last_channel.clone(),
                cache_last_time.clone(),
                slice2_out1.clone(),
                constant672_out1.clone(),
                unsqueeze20_out1.clone(),
                constant444_out1.clone(),
                expand12_out1.clone(),
                constant447_out1.clone(),
                unsqueeze22_out1.clone(),
            );
        let (
            concat72_out1,
            add153_out1,
            concat65_out1,
            concat68_out1,
            concat71_out1,
            slice92_out1,
            slice96_out1,
        ) = self
            .submodule24
            .forward(
                add136_out1,
                constant670_out1.clone(),
                cache_last_channel,
                cache_last_time,
                slice2_out1,
                constant672_out1,
                unsqueeze20_out1,
                constant444_out1,
                expand12_out1,
                constant447_out1,
                unsqueeze22_out1,
            );
        let (slice102_out1, clip3_out1, slice101_out1, concat74_out1, clip2_out1) = self
            .submodule25
            .forward(
                concat72_out1,
                add153_out1,
                constant670_out1,
                concat2_out1,
                concat5_out1,
                concat8_out1,
                concat11_out1,
                concat14_out1,
                concat17_out1,
                concat20_out1,
                concat23_out1,
                concat26_out1,
                concat29_out1,
                concat32_out1,
                concat35_out1,
                concat38_out1,
                concat41_out1,
                concat44_out1,
                concat47_out1,
                concat50_out1,
                concat53_out1,
                concat56_out1,
                concat59_out1,
                concat62_out1,
                concat65_out1,
                concat68_out1,
                concat71_out1,
                slice8_out1,
                slice12_out1,
                slice16_out1,
                slice20_out1,
                slice24_out1,
                slice28_out1,
                slice32_out1,
                slice36_out1,
                slice40_out1,
                slice44_out1,
                slice48_out1,
                slice52_out1,
                slice56_out1,
                slice60_out1,
                slice64_out1,
                slice68_out1,
                slice72_out1,
                slice76_out1,
                slice80_out1,
                slice84_out1,
                slice88_out1,
                slice92_out1,
                slice96_out1,
                cache_last_channel_len,
                clip1_out1,
            );
        (slice102_out1, clip3_out1, slice101_out1, concat74_out1, clip2_out1)
    }
}