strisunshinewentingwang/pytorch

Pull Requests Commits

Workaround Python3.9 limitations in test_jit_py3 (#51088)

malfet committed 3 years ago

a949d7b1

Move AcceleratedGraphModule out of graph_manipulation.

gcatron committed 3 years ago

c8a24ebe

Revert D26018916: [pytorch][PR] Automated submodule update: tensorpipe

izdeby committed 3 years ago

81ae8edf

[quant][graphmode][fx] cleanup linear module test case (#50976)

jerryzh168 committed 3 years ago

afa79a4d

Enable BFloat support for gemms on arch other than ampere (#50442)

zasdfgbnm committed 3 years ago

b822aba8

[dist_optim] add warning to distributed optimizer (#50630)

wanchaol committed 3 years ago

3562ca2d

[reland] Refactor mypy configs list into editor-friendly wrapper (#50826)

samestep committed 3 years ago

6dda0363

[jit] Fix ResolutionCallback definition (#51089)

malfet committed 3 years ago

31194750

Fix test_jit_cuda_archflags on machine with more than one arch (#50405)

zasdfgbnm committed 3 years ago

5834b3b2

Automated submodule update: tensorpipe (#50946)

facebook-github-bot committed 3 years ago

5f297cc6

Enable ROCM Skipped tests in test_ops.py (#50500)

arindamroy-eng committed 3 years ago

95ae9a20

Implement autograd functions for c10d communication operations (#40762)

Emilio Castillo committed 3 years ago

233e4ebd

Turn on batched grad testing for CriterionTest (#50744)

zou3519 committed 3 years ago

83315965

Revert D25850783: Add torch::deploy, an embedded torch-python interpreter

Mike Ruberry committed 3 years ago

e843974a

Improve docs around Math/DefaultBackend & add PythonDispatcher class. (#50854)

Ailing Zhang committed 3 years ago

a51b9a82

[Gradient Compression] Typo fixes in PowerSGD (#50974)

Yi Wang committed 3 years ago

9f19843d

[Gradient Compression] Allow PowerSGD to run vallina allreduce for the first K iterations (#50973)

Yi Wang committed 3 years ago

ffaae32d

Add torch.eig complex forward (CPU, CUDA) (#49168)

antocuni committed 3 years ago

880f0074

Added cuda bindings for NNC (#51046)

Chillee committed 3 years ago

502ca010

[PT QNNPACK] Temporarily disable input pointer caching (#51051)

kimishpatel committed 3 years ago

6ef66213

Abort node in fast_nvcc if ancestor fails (#51043)

samestep committed 3 years ago

5adbace8

Fix TransformedDistribution shaping logic (#50581)

fritzo committed 3 years ago

a347c747

Create a DDPLoggingData and expose it to python interface (#50622)

zhaojuanmao committed 3 years ago

250c7112

Add torch::deploy, an embedded torch-python interpreter (#50458)

wconstab committed 3 years ago

3192f9e4

Make torch.svd return V, not V.conj() for complex inputs (#51012)

IvanYashchuk committed 3 years ago

ddf26816

fake_quant: fix device affinity and buffer resizing for state_dict (#50868)

vkuzo committed 3 years ago

f8eefbdf

Add documentation page for pipeline parallelism. (#50791)

pritamdamania committed 3 years ago

68c21854

Workaround for MAGMA accessing illegal memory in batched cholesky (#50957)

heitorschueroff committed 3 years ago

a7cf04ec

Add type annotations to torch.overrides (#50824)

guilhermeleobas committed 3 years ago

9dfbfe9f

More about cudnn refactor (#50827)

zasdfgbnm committed 3 years ago

75cba9d0

Older