Skip to content

ROI pool CUDA: compare in acc_type to fix half/MSVC build #9399

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

Open
kimchioverfit wants to merge 6 commits into
base: main
Choose a base branch
from
Open

ROI pool CUDA: compare in acc_type to fix half/MSVC build #9399

Changes from 4 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
13ba932
ROI pool CUDA: compare in acc_type to fix half/MSVC build
kimchioverfit Feb 21, 2026
c03784b
Add AccumulateType header to roi_pool_kernel.cu
kimchioverfit Feb 23, 2026
788bf6a
Apply clang-format to roi_pool_kernel.cu
kimchioverfit Feb 24, 2026
3c4b6e0
Merge branch 'main' into fix/roi-pool-half-msvc
kimchioverfit Feb 24, 2026
193825d
Merge branch 'main' into fix/roi-pool-half-msvc
NicolasHug Feb 25, 2026
1d9444a
Merge branch 'main' into fix/roi-pool-half-msvc
kimchioverfit Feb 27, 2026
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
7 changes: 6 additions & 1 deletion torchvision/csrc/ops/cuda/roi_pool_kernel.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,5 @@
#include <ATen/ATen.h>
#include <ATen/AccumulateType.h>
#include <ATen/cuda/CUDAContext.h>
#include <c10/cuda/CUDAGuard.h>
#include <float.h>
Expand Down Expand Up @@ -63,10 +64,14 @@ __global__ void roi_pool_forward_kernel_impl(
int maxidx = -1;
const T* offset_input =
input + (roi_batch_ind * channels + c) * height * width;
using acc_t = at::acc_type<T, /*is_cuda=*/true>;
for (int h = hstart; h < hend; ++h) {
for (int w = wstart; w < wend; ++w) {
int input_index = h * width + w;
if (offset_input[input_index] > maxval) {
acc_t v = static_cast<acc_t>(offset_input[input_index]);
acc_t mv = static_cast<acc_t>(maxval);

if (v > mv) {
maxval = offset_input[input_index];
maxidx = input_index;
}
Expand Down
Loading