pytorch/c10
Nikita Shulga 3525b834f0 [MPSInductor] Implement argmax/argmin (#146429)
TODOs:
 - Find test with NaN
 - Report internal compiler error when running `test_argmax_argmin1` (which is actually not enough shared memory)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/146429
Approved by: https://github.com/dcci
ghstack dependencies: #146423, #146428
2025-02-04 19:16:06 +00:00
..
benchmark Set RUNPATH so installed tests can find the required shared libraries (#136627) 2024-10-25 09:38:08 +00:00
core Use std::string_view (#145906) 2025-01-30 03:14:27 +00:00
cuda use copy2d in h2d/d2h copy when possible (#146256) 2025-02-03 23:07:54 +00:00
hip Fix hardcoded ROCm paths in Caffe2Targets.cmake (#136283) 2024-09-26 00:34:43 +00:00
macros [ROCm][Windows] Fix export macros (#144098) 2025-01-04 17:12:46 +00:00
metal [MPSInductor] Implement argmax/argmin (#146429) 2025-02-04 19:16:06 +00:00
mobile [2/N] Fix extra warnings brought by clang-tidy-17 (#137459) 2024-10-08 19:05:02 +00:00
test Fix cppcoreguidelines-init-variables ignorance (#141795) 2025-01-28 17:11:37 +00:00
util initialize device when pinning memory on this device, short circuit i… (#145752) 2025-01-30 21:37:29 +00:00
xpu Filter out iGPU if dGPU is found on XPU (#144378) 2025-01-29 15:53:16 +00:00
BUCK.oss
BUILD.bazel
build.bzl
CMakeLists.txt [pytorch][monitoring] Dynamic backend for WaitCounter (#135967) 2024-09-15 18:07:49 +00:00
ovrsource_defs.bzl [NeuralNetInference] Bring up iOS builds (#131917) 2024-07-30 23:01:09 +00:00