dfdx 0.13.0

Ergonomic auto differentiation in Rust, with pytorch like apis.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56

use super::{BinarySubKernelOp as Binary, ScalarSubKernelOp as Scalar};
#[allow(unused_imports)]
use crate::dtypes::*;
use crate::tensor_ops::cuda_kernels::{cuda_binary, cuda_unary};

#[cfg(feature = "f16")]
unsafe impl cudarc::driver::DeviceRepr for Scalar<f16> {}
#[cfg(feature = "f16")]
unsafe impl cudarc::driver::DeviceRepr for Scalar<AMP<f16>> {}
unsafe impl cudarc::driver::DeviceRepr for Scalar<f32> {}
unsafe impl cudarc::driver::DeviceRepr for Scalar<f64> {}
unsafe impl cudarc::driver::DeviceRepr for Binary {}

const SCALAR_PTX: &str = include_str!(concat!(env!("OUT_DIR"), "/scalar_sub.ptx"));
const BINARY_PTX: &str = include_str!(concat!(env!("OUT_DIR"), "/binary_sub.ptx"));

#[cfg(feature = "f16")]
cuda_unary!(const_df() Scalar<f16>, f16, SCALAR_PTX, "ssub_fwd_f16", "ssub_bwd_f16");
#[cfg(feature = "f16")]
cuda_unary!(const_df() Scalar<AMP<f16>>, AMP<f16>, SCALAR_PTX, "ssub_fwd_f16", "ssub_bwd_f16");
cuda_unary!(const_df() Scalar<f32>, f32, SCALAR_PTX, "ssub_fwd_f32", "ssub_bwd_f32");
cuda_unary!(const_df() Scalar<f64>, f64, SCALAR_PTX, "ssub_fwd_f64", "ssub_bwd_f64");
#[cfg(feature = "f16")]
cuda_binary!(
    const_df() Binary,
    f16,
    BINARY_PTX,
    "bsub_fwd_f16",
    "bsub_bwd_lhs_f16",
    "bsub_bwd_rhs_f16"
);
#[cfg(feature = "f16")]
cuda_binary!(
    const_df() Binary,
    AMP<f16>,
    BINARY_PTX,
    "bsub_fwd_f16",
    "bsub_bwd_lhs_f16",
    "bsub_bwd_rhs_f16"
);
cuda_binary!(
    const_df() Binary,
    f32,
    BINARY_PTX,
    "bsub_fwd_f32",
    "bsub_bwd_lhs_f32",
    "bsub_bwd_rhs_f32"
);
cuda_binary!(
    const_df() Binary,
    f64,
    BINARY_PTX,
    "bsub_fwd_f64",
    "bsub_bwd_lhs_f64",
    "bsub_bwd_rhs_f64"
);