Trait ModuleOps 

pub trait ModuleOps<B: Backend> {
    // Required methods
    fn conv2d(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        bias: Option<FloatTensor<B>>,
        options: ConvOptions<2>,
    ) -> FloatTensor<B>;
    fn deform_conv2d(
        x: FloatTensor<B>,
        offset: FloatTensor<B>,
        weight: FloatTensor<B>,
        mask: Option<FloatTensor<B>>,
        bias: Option<FloatTensor<B>>,
        options: DeformConvOptions<2>,
    ) -> FloatTensor<B>;
    fn deform_conv2d_backward(
        x: FloatTensor<B>,
        offset: FloatTensor<B>,
        weight: FloatTensor<B>,
        mask: Option<FloatTensor<B>>,
        bias: Option<FloatTensor<B>>,
        output_grad: FloatTensor<B>,
        options: DeformConvOptions<2>,
    ) -> DeformConv2dBackward<B>;
    fn conv3d(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        bias: Option<FloatTensor<B>>,
        options: ConvOptions<3>,
    ) -> FloatTensor<B>;
    fn conv_transpose2d(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        bias: Option<FloatTensor<B>>,
        options: ConvTransposeOptions<2>,
    ) -> FloatTensor<B>;
    fn conv_transpose3d(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        bias: Option<FloatTensor<B>>,
        options: ConvTransposeOptions<3>,
    ) -> FloatTensor<B>;
    fn avg_pool2d(
        x: FloatTensor<B>,
        kernel_size: [usize; 2],
        stride: [usize; 2],
        padding: [usize; 2],
        count_include_pad: bool,
    ) -> FloatTensor<B>;
    fn avg_pool2d_backward(
        x: FloatTensor<B>,
        grad: FloatTensor<B>,
        kernel_size: [usize; 2],
        stride: [usize; 2],
        padding: [usize; 2],
        count_include_pad: bool,
    ) -> FloatTensor<B>;
    fn adaptive_avg_pool2d(
        x: FloatTensor<B>,
        output_size: [usize; 2],
    ) -> FloatTensor<B>;
    fn adaptive_avg_pool2d_backward(
        x: FloatTensor<B>,
        grad: FloatTensor<B>,
    ) -> FloatTensor<B>;
    fn max_pool2d(
        x: FloatTensor<B>,
        kernel_size: [usize; 2],
        stride: [usize; 2],
        padding: [usize; 2],
        dilation: [usize; 2],
    ) -> FloatTensor<B>;
    fn max_pool2d_with_indices(
        x: FloatTensor<B>,
        kernel_size: [usize; 2],
        stride: [usize; 2],
        padding: [usize; 2],
        dilation: [usize; 2],
    ) -> MaxPool2dWithIndices<B>;
    fn max_pool2d_with_indices_backward(
        x: FloatTensor<B>,
        kernel_size: [usize; 2],
        stride: [usize; 2],
        padding: [usize; 2],
        dilation: [usize; 2],
        output_grad: FloatTensor<B>,
        indices: IntTensor<B>,
    ) -> MaxPool2dBackward<B>;
    fn interpolate(
        x: FloatTensor<B>,
        output_size: [usize; 2],
        options: InterpolateOptions,
    ) -> FloatTensor<B>;
    fn interpolate_backward(
        x: FloatTensor<B>,
        grad: FloatTensor<B>,
        output_size: [usize; 2],
        options: InterpolateOptions,
    ) -> FloatTensor<B>;

    // Provided methods
    fn embedding(
        weights: FloatTensor<B>,
        indices: IntTensor<B>,
    ) -> FloatTensor<B> { ... }
    fn embedding_backward(
        weights: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        indices: IntTensor<B>,
    ) -> FloatTensor<B> { ... }
    fn conv1d(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        bias: Option<FloatTensor<B>>,
        options: ConvOptions<1>,
    ) -> FloatTensor<B> { ... }
    fn conv1d_x_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvOptions<1>,
    ) -> FloatTensor<B> { ... }
    fn conv1d_weight_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvOptions<1>,
    ) -> FloatTensor<B> { ... }
    fn conv1d_bias_backward(
        x: FloatTensor<B>,
        bias: FloatTensor<B>,
        output_grad: FloatTensor<B>,
    ) -> FloatTensor<B> { ... }
    fn conv2d_x_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvOptions<2>,
    ) -> FloatTensor<B> { ... }
    fn conv2d_weight_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvOptions<2>,
    ) -> FloatTensor<B> { ... }
    fn conv2d_bias_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        bias: FloatTensor<B>,
        output_grad: FloatTensor<B>,
    ) -> FloatTensor<B> { ... }
    fn conv3d_x_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvOptions<3>,
    ) -> FloatTensor<B> { ... }
    fn conv3d_weight_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvOptions<3>,
    ) -> FloatTensor<B> { ... }
    fn conv3d_bias_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        bias: FloatTensor<B>,
        output_grad: FloatTensor<B>,
    ) -> FloatTensor<B> { ... }
    fn conv_transpose1d(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        bias: Option<FloatTensor<B>>,
        options: ConvTransposeOptions<1>,
    ) -> FloatTensor<B> { ... }
    fn conv_transpose1d_x_backward(
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvTransposeOptions<1>,
    ) -> FloatTensor<B> { ... }
    fn conv_transpose1d_weight_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvTransposeOptions<1>,
    ) -> FloatTensor<B> { ... }
    fn conv_transpose1d_bias_backward(
        x: FloatTensor<B>,
        bias: FloatTensor<B>,
        output_grad: FloatTensor<B>,
    ) -> FloatTensor<B> { ... }
    fn conv_transpose2d_x_backward(
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvTransposeOptions<2>,
    ) -> FloatTensor<B> { ... }
    fn conv_transpose2d_weight_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvTransposeOptions<2>,
    ) -> FloatTensor<B> { ... }
    fn conv_transpose2d_bias_backward(
        x: FloatTensor<B>,
        bias: FloatTensor<B>,
        output_grad: FloatTensor<B>,
    ) -> FloatTensor<B> { ... }
    fn conv_transpose3d_x_backward(
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvTransposeOptions<3>,
    ) -> FloatTensor<B> { ... }
    fn conv_transpose3d_weight_backward(
        x: FloatTensor<B>,
        weight: FloatTensor<B>,
        output_grad: FloatTensor<B>,
        options: ConvTransposeOptions<3>,
    ) -> FloatTensor<B> { ... }
    fn conv_transpose3d_bias_backward(
        x: FloatTensor<B>,
        bias: FloatTensor<B>,
        output_grad: FloatTensor<B>,
    ) -> FloatTensor<B> { ... }
    fn unfold4d(
        x: FloatTensor<B>,
        kernel_size: [usize; 2],
        options: UnfoldOptions,
    ) -> FloatTensor<B> { ... }
    fn avg_pool1d(
        x: FloatTensor<B>,
        kernel_size: usize,
        stride: usize,
        padding: usize,
        count_include_pad: bool,
    ) -> FloatTensor<B> { ... }
    fn avg_pool1d_backward(
        x: FloatTensor<B>,
        grad: FloatTensor<B>,
        kernel_size: usize,
        stride: usize,
        padding: usize,
        count_include_pad: bool,
    ) -> FloatTensor<B> { ... }
    fn adaptive_avg_pool1d(
        x: FloatTensor<B>,
        output_size: usize,
    ) -> FloatTensor<B> { ... }
    fn adaptive_avg_pool1d_backward(
        x: FloatTensor<B>,
        grad: FloatTensor<B>,
    ) -> FloatTensor<B> { ... }
    fn max_pool1d(
        x: FloatTensor<B>,
        kernel_size: usize,
        stride: usize,
        padding: usize,
        dilation: usize,
    ) -> FloatTensor<B> { ... }
    fn max_pool1d_with_indices(
        x: FloatTensor<B>,
        kernel_size: usize,
        stride: usize,
        padding: usize,
        dilation: usize,
    ) -> MaxPool1dWithIndices<B> { ... }
    fn max_pool1d_with_indices_backward(
        x: FloatTensor<B>,
        kernel_size: usize,
        stride: usize,
        padding: usize,
        dilation: usize,
        output_grad: FloatTensor<B>,
        indices: IntTensor<B>,
    ) -> MaxPool1dBackward<B> { ... }
}

Module operations trait.

Required Methods

fn conv2d( x: FloatTensor<B>, weight: FloatTensor<B>, bias: Option<FloatTensor<B>>, options: ConvOptions<2>, ) -> FloatTensor<B>

Two dimensional convolution.

Shapes

x: [batch_size, channels_in, height, width], weight: [channels_out, channels_in, kernel_size_1, kernel_size_2], bias: [channels_out],

fn deform_conv2d( x: FloatTensor<B>, offset: FloatTensor<B>, weight: FloatTensor<B>, mask: Option<FloatTensor<B>>, bias: Option<FloatTensor<B>>, options: DeformConvOptions<2>, ) -> FloatTensor<B>

Two dimensional deformable convolution.

Shapes

x: [batch_size, channels_in, height, width], weight: [channels_out, channels_in, kernel_size_1, kernel_size_2], bias: [channels_out],

fn deform_conv2d_backward( x: FloatTensor<B>, offset: FloatTensor<B>, weight: FloatTensor<B>, mask: Option<FloatTensor<B>>, bias: Option<FloatTensor<B>>, output_grad: FloatTensor<B>, options: DeformConvOptions<2>, ) -> DeformConv2dBackward<B>

Backward pass for the deform_conv2d operation.

fn conv3d( x: FloatTensor<B>, weight: FloatTensor<B>, bias: Option<FloatTensor<B>>, options: ConvOptions<3>, ) -> FloatTensor<B>

Three dimensional convolution.

Shapes

x: [batch_size, channels_in, depth, height, width], weight: [channels_out, channels_in, kernel_size_1, kernel_size_2, kernel_size_3], bias: [channels_out],

fn conv_transpose2d( x: FloatTensor<B>, weight: FloatTensor<B>, bias: Option<FloatTensor<B>>, options: ConvTransposeOptions<2>, ) -> FloatTensor<B>

Two dimensional transposed convolution.

Shapes

x: [batch_size, channels_in, height, width], weight: [channels_in, channels_out, kernel_size_1, kernel_size_2], bias: [channels_out],

fn conv_transpose3d( x: FloatTensor<B>, weight: FloatTensor<B>, bias: Option<FloatTensor<B>>, options: ConvTransposeOptions<3>, ) -> FloatTensor<B>

Three dimensional transposed convolution.

Shapes

x: [batch_size, channels_in, height, width], weight: [channels_in, channels_out, kernel_size_1, kernel_size_2, kernel_size_3], bias: [channels_out],

fn avg_pool2d( x: FloatTensor<B>, kernel_size: [usize; 2], stride: [usize; 2], padding: [usize; 2], count_include_pad: bool, ) -> FloatTensor<B>

Two dimensional avg pooling.

Shapes

x: [batch_size, channels, height, width],

fn avg_pool2d_backward( x: FloatTensor<B>, grad: FloatTensor<B>, kernel_size: [usize; 2], stride: [usize; 2], padding: [usize; 2], count_include_pad: bool, ) -> FloatTensor<B>

Backward pass for the avg pooling 2d operation.

fn adaptive_avg_pool2d( x: FloatTensor<B>, output_size: [usize; 2], ) -> FloatTensor<B>

Two dimensional adaptive avg pooling.

Shapes

x: [batch_size, channels, height, width],

fn adaptive_avg_pool2d_backward( x: FloatTensor<B>, grad: FloatTensor<B>, ) -> FloatTensor<B>

Backward pass for the adaptive avg pooling 2d operation.

fn max_pool2d( x: FloatTensor<B>, kernel_size: [usize; 2], stride: [usize; 2], padding: [usize; 2], dilation: [usize; 2], ) -> FloatTensor<B>

Two dimensional max pooling.

Shapes

x: [batch_size, channels, height, width],

fn max_pool2d_with_indices( x: FloatTensor<B>, kernel_size: [usize; 2], stride: [usize; 2], padding: [usize; 2], dilation: [usize; 2], ) -> MaxPool2dWithIndices<B>

Two dimensional max pooling with indices.

Shapes

x: [batch_size, channels, height, width],

fn max_pool2d_with_indices_backward( x: FloatTensor<B>, kernel_size: [usize; 2], stride: [usize; 2], padding: [usize; 2], dilation: [usize; 2], output_grad: FloatTensor<B>, indices: IntTensor<B>, ) -> MaxPool2dBackward<B>

Backward pass for the max pooling 2d operation.

fn interpolate( x: FloatTensor<B>, output_size: [usize; 2], options: InterpolateOptions, ) -> FloatTensor<B>

Down/up samples the input.

Shapes

x: [batch_size, channels, height, width],

fn interpolate_backward( x: FloatTensor<B>, grad: FloatTensor<B>, output_size: [usize; 2], options: InterpolateOptions, ) -> FloatTensor<B>

Backward pass for the interpolate operation.

Provided Methods

fn embedding(weights: FloatTensor<B>, indices: IntTensor<B>) -> FloatTensor<B>

Embedding operation.

Arguments

• weights - The embedding weights.
• indices - The indices tensor.

Returns

The output tensor.

fn embedding_backward( weights: FloatTensor<B>, output_grad: FloatTensor<B>, indices: IntTensor<B>, ) -> FloatTensor<B>

Embedding backward operation.

Arguments

• weights - The embedding weights.
• output_grad - The output gradient.
• indices - The indices tensor.

Returns

The gradient.

fn conv1d( x: FloatTensor<B>, weight: FloatTensor<B>, bias: Option<FloatTensor<B>>, options: ConvOptions<1>, ) -> FloatTensor<B>

One dimensional convolution.

Shapes

x: [batch_size, channels_in, length], weight: [channels_out, channels_in, kernel_size], bias: [channels_out],

fn conv1d_x_backward( x: FloatTensor<B>, weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvOptions<1>, ) -> FloatTensor<B>

Backward pass for the conv1d operation, returning the gradient for x.

fn conv1d_weight_backward( x: FloatTensor<B>, weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvOptions<1>, ) -> FloatTensor<B>

Backward pass for the conv1d operation, returning the gradient for weight.

fn conv1d_bias_backward( x: FloatTensor<B>, bias: FloatTensor<B>, output_grad: FloatTensor<B>, ) -> FloatTensor<B>

Backward pass for the conv1d operation, returning the gradient for bias.

fn conv2d_x_backward( x: FloatTensor<B>, weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvOptions<2>, ) -> FloatTensor<B>

Backward pass for the conv2d operation, returning the gradient for x.

fn conv2d_weight_backward( x: FloatTensor<B>, weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvOptions<2>, ) -> FloatTensor<B>

Backward pass for the conv2d operation, returning the gradient for weight.

fn conv2d_bias_backward( x: FloatTensor<B>, weight: FloatTensor<B>, bias: FloatTensor<B>, output_grad: FloatTensor<B>, ) -> FloatTensor<B>

Backward pass for the conv2d operation, returning the gradient for bias.

fn conv3d_x_backward( x: FloatTensor<B>, weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvOptions<3>, ) -> FloatTensor<B>

Backward pass for the conv3d operation, returning the gradient for x.

fn conv3d_weight_backward( x: FloatTensor<B>, weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvOptions<3>, ) -> FloatTensor<B>

Backward pass for the conv3d operation, returning the gradient for weight.

fn conv3d_bias_backward( x: FloatTensor<B>, weight: FloatTensor<B>, bias: FloatTensor<B>, output_grad: FloatTensor<B>, ) -> FloatTensor<B>

Backward pass for the conv3d operation, returning the gradient for bias.

fn conv_transpose1d( x: FloatTensor<B>, weight: FloatTensor<B>, bias: Option<FloatTensor<B>>, options: ConvTransposeOptions<1>, ) -> FloatTensor<B>

One dimensional transposed convolution.

Shapes

x: [batch_size, channels_in, length], weight: [channels_in, channels_out, length], bias: [channels_out],

fn conv_transpose1d_x_backward( weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvTransposeOptions<1>, ) -> FloatTensor<B>

Backward pass for the conv transpose 1d operation, returning the gradient for x.

fn conv_transpose1d_weight_backward( x: FloatTensor<B>, weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvTransposeOptions<1>, ) -> FloatTensor<B>

Backward pass for the conv transpose 1d operation, returning the gradient for weight.

fn conv_transpose1d_bias_backward( x: FloatTensor<B>, bias: FloatTensor<B>, output_grad: FloatTensor<B>, ) -> FloatTensor<B>

Backward pass for the conv transpose 1d operation, returning the gradient for bias.

fn conv_transpose2d_x_backward( weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvTransposeOptions<2>, ) -> FloatTensor<B>

Backward pass for the conv transpose 2d operation, returning the gradient for x.

fn conv_transpose2d_weight_backward( x: FloatTensor<B>, weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvTransposeOptions<2>, ) -> FloatTensor<B>

Backward pass for the conv transpose 2d operation, returning the gradient for weight.

fn conv_transpose2d_bias_backward( x: FloatTensor<B>, bias: FloatTensor<B>, output_grad: FloatTensor<B>, ) -> FloatTensor<B>

Backward pass for the conv transpose 2d operation, returning the gradient for bias.

fn conv_transpose3d_x_backward( weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvTransposeOptions<3>, ) -> FloatTensor<B>

Backward pass for the conv transpose 3d operation, returning the gradient for x.

fn conv_transpose3d_weight_backward( x: FloatTensor<B>, weight: FloatTensor<B>, output_grad: FloatTensor<B>, options: ConvTransposeOptions<3>, ) -> FloatTensor<B>

Backward pass for the conv transpose 3d operation, returning the gradient for weight.

fn conv_transpose3d_bias_backward( x: FloatTensor<B>, bias: FloatTensor<B>, output_grad: FloatTensor<B>, ) -> FloatTensor<B>

Backward pass for the conv transpose 3d operation, returning the gradient for bias.

fn unfold4d( x: FloatTensor<B>, kernel_size: [usize; 2], options: UnfoldOptions, ) -> FloatTensor<B>

Four-dimensional unfolding.

Shapes

• x: [batch_size, channels_in, height, width],
• returns: [batch_size, channels_in * kernel_size_1 * kernel_size_2, number of blocks],

fn avg_pool1d( x: FloatTensor<B>, kernel_size: usize, stride: usize, padding: usize, count_include_pad: bool, ) -> FloatTensor<B>

One dimensional avg pooling.

Shapes

x: [batch_size, channels, length],

fn avg_pool1d_backward( x: FloatTensor<B>, grad: FloatTensor<B>, kernel_size: usize, stride: usize, padding: usize, count_include_pad: bool, ) -> FloatTensor<B>

Backward pass for the avg pooling 1d operation.

fn adaptive_avg_pool1d(x: FloatTensor<B>, output_size: usize) -> FloatTensor<B>

One dimensional adaptive avg pooling.

Shapes

x: [batch_size, channels, length],

fn adaptive_avg_pool1d_backward( x: FloatTensor<B>, grad: FloatTensor<B>, ) -> FloatTensor<B>

Backward pass for the adaptive avg pooling 1d operation.

fn max_pool1d( x: FloatTensor<B>, kernel_size: usize, stride: usize, padding: usize, dilation: usize, ) -> FloatTensor<B>

One dimensional max pooling.

Shapes

x: [batch_size, channels, length],

fn max_pool1d_with_indices( x: FloatTensor<B>, kernel_size: usize, stride: usize, padding: usize, dilation: usize, ) -> MaxPool1dWithIndices<B>

One dimensional max pooling with indices.

Shapes

x: [batch_size, channels, height, width],

fn max_pool1d_with_indices_backward( x: FloatTensor<B>, kernel_size: usize, stride: usize, padding: usize, dilation: usize, output_grad: FloatTensor<B>, indices: IntTensor<B>, ) -> MaxPool1dBackward<B>

Backward pass for the max pooling 1d operation.

Dyn Compatibility

This trait is not dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety", so this trait is not object safe.

Implementors