Enable im2col on the cpu side. (#805)

* Enable im2col on the cpu side.

* Hook im2col on the cpu backend.

* Use the kernel offset.

* Avoid an unnecessary copy.

* Handle non-contiguous kernels.

* Add a const to select the conv2d kernel.
This commit is contained in:
Laurent Mazare
2023-09-11 09:28:13 +01:00
committed by GitHub
parent 1cd74129d4
commit 6fb665004c

View File

@ -4,6 +4,8 @@ use crate::{DType, Error, IntDType, Layout, Result, Shape, WithDType};
use half::{bf16, f16}; use half::{bf16, f16};
use rayon::prelude::*; use rayon::prelude::*;
const USE_IM2COL_CONV2D: bool = true;
// TODO: Maybe we should not implement [Clone] here and instead have an explicit allocator + // TODO: Maybe we should not implement [Clone] here and instead have an explicit allocator +
// intercept the oom errors to avoid panicking and provide a proper error. // intercept the oom errors to avoid panicking and provide a proper error.
#[derive(Debug, Clone)] #[derive(Debug, Clone)]
@ -1089,6 +1091,81 @@ impl<'a> Map2 for Conv1D<'a> {
} }
} }
struct Im2Col {
h_k: usize,
w_k: usize,
stride: usize,
dilation: usize,
padding: usize,
}
impl Im2Col {
fn hw_out(&self, h: usize, w: usize) -> (usize, usize) {
let h_out = (h + 2 * self.padding - self.dilation * (self.h_k - 1) - 1) / self.stride + 1;
let w_out = (w + 2 * self.padding - self.dilation * (self.w_k - 1) - 1) / self.stride + 1;
(h_out, w_out)
}
}
impl Map1 for Im2Col {
fn f<T: WithDType>(&self, vs: &[T], layout: &Layout) -> Result<Vec<T>> {
let &Self {
h_k,
w_k,
stride,
dilation,
padding,
} = self;
let (b, c, h, w) = layout.shape().dims4()?;
let (h_out, w_out) = self.hw_out(h, w);
let src = &vs[layout.start_offset()..];
let mut dst = vec![T::zero(); b * h_out * w_out * c * h_k * w_k];
let (src_s0, src_s1, src_s2, src_s3) = {
let s = layout.stride();
(s[0], s[1], s[2], s[3])
};
// TODO: provide specialized kernels for the common use cases.
// - h_k = w_k = 1
// - padding = 0
// - stride = 1
// - dilation = 1
for b_idx in 0..b {
let src_idx = b_idx * src_s0;
let dst_idx = b_idx * h_out * w_out * c * h_k * w_k;
for h_idx in 0..h_out {
let dst_idx = dst_idx + h_idx * w_out * c * h_k * w_k;
for w_idx in 0..w_out {
let dst_idx = dst_idx + w_idx * c * h_k * w_k;
for c_idx in 0..c {
let dst_idx = dst_idx + c_idx * h_k * w_k;
let src_idx = c_idx * src_s1 + src_idx;
for h_k_idx in 0..h_k {
let src_h = h_idx * stride + h_k_idx * dilation;
if padding != 0 && (src_h < padding || src_h >= h + padding) {
continue;
}
let src_h = src_h - padding;
let src_idx = src_idx + src_h * src_s2;
let dst_idx = dst_idx + h_k_idx * w_k;
for w_k_idx in 0..w_k {
let src_w = w_idx * stride + w_k_idx * dilation;
if padding != 0 && (src_w < padding || src_w >= h + padding) {
continue;
}
let src_w = src_w - padding;
let src_idx = src_idx + src_w * src_s3;
let dst_idx = dst_idx + w_k_idx;
dst[dst_idx] = src[src_idx]
}
}
}
}
}
}
Ok(dst)
}
}
struct Conv2D<'a>(&'a crate::conv::ParamsConv2D); struct Conv2D<'a>(&'a crate::conv::ParamsConv2D);
impl<'a> Map2 for Conv2D<'a> { impl<'a> Map2 for Conv2D<'a> {
@ -2237,7 +2314,43 @@ impl BackendStorage for CpuStorage {
kernel_l: &Layout, kernel_l: &Layout,
params: &crate::conv::ParamsConv2D, params: &crate::conv::ParamsConv2D,
) -> Result<Self> { ) -> Result<Self> {
Conv2D(params).map(self, l, kernel, kernel_l) if !USE_IM2COL_CONV2D {
return Conv2D(params).map(self, l, kernel, kernel_l);
}
let op = Im2Col {
h_k: params.k_h,
w_k: params.k_w,
padding: params.padding,
stride: params.stride,
dilation: params.dilation,
};
let col = op.map(self, l)?;
let b = params.b_size;
let n = params.c_out;
let (h_out, w_out) = (params.out_h(), params.out_w());
let k = op.h_k * op.w_k * params.c_in;
let m = h_out * w_out;
let col_l = Layout::contiguous((b, m, k));
let res = if kernel_l.is_contiguous() {
let kernel_l = Layout::contiguous_with_offset((1, n, k), kernel_l.start_offset())
.transpose(1, 2)?
.broadcast_as((b, k, n))?;
col.matmul(kernel, (b, m, n, k), &col_l, &kernel_l)?
} else {
// Make the kernel contiguous if not already the case.
let mut kernel_c = self.device().zeros_impl(kernel_l.shape(), kernel.dtype())?;
kernel.copy_strided_src(&mut kernel_c, 0, kernel_l)?;
let kernel_l = Layout::contiguous_with_offset((1, n, k), kernel_l.start_offset())
.transpose(1, 2)?
.broadcast_as((b, k, n))?;
col.matmul(kernel, (b, m, n, k), &col_l, &kernel_l)?
};
let res_l = Layout::contiguous((b, h_out, w_out, params.c_out))
.transpose(1, 2)?
.transpose(1, 3)?;
let mut res_t = self.device().zeros_impl(res_l.shape(), res.dtype())?;
res.copy_strided_src(&mut res_t, 0, &res_l)?;
Ok(res_t)
} }
fn conv_transpose2d( fn conv_transpose2d(