Skip to content

fix im2col_f32_f16 #658

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Dec 18, 2023
Merged

fix im2col_f32_f16 #658

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 4 additions & 4 deletions src/ggml-cuda.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5259,17 +5259,17 @@ static __global__ void im2col_f32_f16(
const int ky = (i - kd) / OW;
const int ix = i % OW;

const int iiw = ix * s0 + kx * d0 - p0;
const int iih = blockIdx.y * s1 + ky * d1 - p1;
const int64_t iiw = ix * s0 + kx * d0 - p0;
const int64_t iih = blockIdx.y * s1 + ky * d1 - p1;

const int offset_dst =
const int64_t offset_dst =
(blockIdx.y * OW + ix) * CHW +
(blockIdx.z * (KW * KH) + ky * KW + kx);

if (iih < 0 || iih >= IH || iiw < 0 || iiw >= IW) {
dst[offset_dst] = __float2half(0.0f);
} else {
const int offset_src = blockIdx.z * offset_delta;
const int64_t offset_src = blockIdx.z * offset_delta;
dst[offset_dst] = __float2half(x[offset_src + iih * IW + iiw]);
}
}
Expand Down