fix: ambiguous min in nms_kernel function

[ddelgadovargas-cyber]

I got the following error when building nms_kernel.cu

third_party/py/torchvision/torchvision/csrc/ops/cuda/nms_kernel.cu:190:7: error: no matching function for call to 'min'
  190 |       min(col_blocks, threadsPerBlock),
      |       ^~~
third_party/crosstool/v18/stable/toolchain/lib/clang/google3-trunk/include/__clang_cuda_math.h:201:16: note: candidate function not viable: call to __device__ function from __host__ function
  201 | __DEVICE__ int min(int __a, int __b) { return __nv_min(__a, __b); }
      |                ^
blaze-out/k8-opt-cuda/bin/third_party/gpus/cuda/_virtual_includes/_stage/third_party/gpus/cuda/include/crt/math_functions.hpp:868:38: note: candidate function not viable: call to __device__ function from __host__ function
  868 | __MATH_FUNCTIONS_DECL__ unsigned int min(const unsigned int a, const unsigned int b)
      |                                      ^
blaze-out/k8-opt-cuda/bin/third_party/gpus/cuda/_virtual_includes/_stage/third_party/gpus/cuda/include/crt/math_functions.hpp:873:38: note: candidate function not viable: call to __device__ function from __host__ function
  873 | __MATH_FUNCTIONS_DECL__ unsigned int min(const int a, const unsigned int b)
      |                                      ^
blaze-out/k8-opt-cuda/bin/third_party/gpus/cuda/_virtual_includes/_stage/third_party/gpus/cuda/include/crt/math_functions.hpp:878:38: note: candidate function not viable: call to __device__ function from __host__ function
  878 | __MATH_FUNCTIONS_DECL__ unsigned int min(const unsigned int a, const int b)
      |                                      ^
note: remaining 12 candidates omitted; pass -fshow-overloads=all to show them

The nms_kernel function is a Host function and the code inside the triple-angle brackets is evaluated on the CPU before the kernel is actually launched on the GPU. So, this change tells the compiler to explicitly use std::min.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/9469

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 9 New Failures, 1 Unrelated Failure

As of commit 33b3ffd with merge base 4e58149 ():

NEW FAILURES - The following jobs have failed:

• Build Windows Wheels / pytorch/vision / build-wheel-py3_10-cuda12_6 (gh)
  ModuleNotFoundError: No module named 'packaging'
• Build Windows Wheels / pytorch/vision / build-wheel-py3_10-cuda12_8 (gh)
  ModuleNotFoundError: No module named 'packaging'
• Build Windows Wheels / pytorch/vision / build-wheel-py3_10-cuda13_0 (gh)
  ModuleNotFoundError: No module named 'packaging'
• Build Windows Wheels / pytorch/vision / build-wheel-py3_10-cuda13_2 (gh)
  RuntimeError: Error compiling objects for extension
• Build Windows Wheels / pytorch/vision / build-wheel-py3_10-xpu (gh)
  ModuleNotFoundError: No module named 'packaging'
• Build Windows Wheels / pytorch/vision / upload / upload-wheel-py3_10-cuda12_6 (gh)
  Unable to download artifact(s): Artifact not found for name: pytorch_vision__3.10_cu126_x64
• Build Windows Wheels / pytorch/vision / upload / upload-wheel-py3_10-cuda13_0 (gh)
  Unable to download artifact(s): Artifact not found for name: pytorch_vision__3.10_cu130_x64
• Build Windows Wheels / pytorch/vision / upload / upload-wheel-py3_10-cuda13_2 (gh)
  Unable to download artifact(s): Artifact not found for name: pytorch_vision__3.10_cu132_x64
• Build Windows Wheels / pytorch/vision / upload / upload-wheel-py3_10-xpu (gh)
  Unable to download artifact(s): Artifact not found for name: pytorch_vision__3.10_xpu_x64

FLAKY - The following job failed but was likely due to flakiness present on trunk:

• Build Windows Wheels / pytorch/vision / upload / upload-wheel-py3_10-cuda12_8 (gh) (detected as infra flaky with no log or failing log classifier)

This comment was automatically generated by Dr. CI and updates every 15 minutes.