strisunshinewentingwang/pytorch

Pull Requests Commits

[ao] Added function to inform dynamic vs static appropriate

vspenubarthi committed 2 years ago

38952d93

Enable torch._refs.var for nvFuser executor (#79517)

IvanYashchuk committed 2 years ago

e10b7625

[Static Runtime] Performance optimization for fork operation (#79482)

akshayParashar1995 committed 2 years ago

20675977

[Vulkan] added vulkan threshold op (#78654)

psavla2 committed 2 years ago

35b130ec

[DataPipe] Fix error message coming from singler iterator constraint

NivekT committed 2 years ago

22c7b1dd

[xla hash update] update the pinned xla hash (#79172)

pytorchmergebot committed 2 years ago

bad7720d

[ci/docs] add some documentation about the stats uploading process

suo committed 2 years ago

6e2f9ece

[ROCm] Enable some sparse tests on ROCm (#77877)

jpvillam-amd committed 2 years ago

aff7eef4

increase sleep for TestCuda.test_caching_pinned_memory_multi_gpu (#76601)

jeffdaily committed 2 years ago

20d56d2b

do not write sccache stats to json if missing OUR_GITHUB_JOB_ID (#79541)

jeffdaily committed 2 years ago

adaafeed

Add support for directly passing symint to empty

ezyang committed 2 years ago

05664a95

Add support for multiply on direct SymInt

ezyang committed 2 years ago

6b015af7

Skip Flaky ONNX Test (#79556)

zengk95 committed 2 years ago

28d1216b

Followup fix after #78828 (#79554)

malfet committed 2 years ago

e895672b

Port linalg_eigh and linalg_eigvalsh to structured

lezcano committed 2 years ago

549a597c

Reference implementations for softmax, log_softmax, logsumexp (#79423)

IvanYashchuk committed 2 years ago

4fc7832d

Enable torch._refs.mean for nvFuser executor (#79444)

IvanYashchuk committed 2 years ago

88958627

Skip extremely long chebyshev/legendre tests introduced in #78304 (#79529)

janeyx99 committed 2 years ago

0005ad88

Revert "Make GHA workflow to retrieve latest promote-able SHA from master (#79548)"

pytorchmergebot committed 2 years ago

dde81c20

[ao] Added ModelReportObserver to inform on dynamic vs static

vspenubarthi committed 2 years ago

8e055131

[DataLoader] Fix the world_size when distributed sharding MapDataPipe (#79524)

ejguan committed 2 years ago

04f87f2a

Make GHA workflow to retrieve latest promote-able SHA from master (#79548)

swang392 committed 2 years ago

e479daed

[const_fold] Set requires_grad based on the folded tensor; add device_for_folding option (#79067)

jfix71 committed 2 years ago

f614f66a

Make flatbuffer loads faster if loading as mobile module. (#78998)

qihqi committed 2 years ago

577f87bb

[MPS] Support stride of stride

malfet committed 2 years ago

81cd276d

Fix warning: cast from type `const char*` to type `char*` casts away qualifiers (#79520)

antoniojkim committed 2 years ago

51b65cd7

Symbolic storage size

ezyang committed 2 years ago

95f9ca49

Add full support for serialization of MPS Tensors (#79465)

albanD committed 2 years ago

0a651a23

[PT-D] Use process group of the partial tensor so sub pg comm will be enabled during reshard

fduwjj committed 2 years ago

f4edbaa6

Revert "Add full support for serialization of MPS Tensors (#79465)"

pytorchmergebot committed 2 years ago

ce6ce747

Newer Older