strisunshinewentingwang/pytorch

Pull Requests Commits

[pytorch] Early return in nn.EmbeddingBag when weight is empty (#46572)

xw285cornell committed 3 years ago

e3b2bfa2

fix-process-group-counter (#46563)

Joel Lamy-Poirier committed 3 years ago

caed29a0

Bump up windows cudnn version (#46436)

zasdfgbnm committed 3 years ago

ce04e527

Workaround to pay attention for CUDA version (#46535)

twdragon committed 3 years ago

c3c249aa

Fix version comparisons for Python 3.6, 3.10 and 4 (#32389)

hugovk committed 3 years ago

09896eda

Apply hip vs hipcc compilation flags correctly for building extensions (#46273)

jithunnair-amd committed 3 years ago

65da50c0

Fix typo in docs for interpolate (#46589)

madebyollin committed 3 years ago

ac4ee0ef

[ONNX] Export var, var_mean and std_mean ops (#45678)

neginraoof committed 3 years ago

96bc7faa

Allow converting parameters of nn.Module to complex dtypes (#44788)

IvanYashchuk committed 3 years ago

6de619e4

Add Batch-Updating Parameter Server Example to CI Tests (#46510)

H-Huang committed 3 years ago

611f0281

first cut of adding a dangling impl test. fix #45165 (#46484)

bdhirsh committed 3 years ago

cf3d7a26

Delete CUDAUnaryOps.cpp (#46280)

zasdfgbnm committed 3 years ago

62e714c9

Revert D24379422: [py][vulkan] Add is_vulkan to py api, add vulkan to device type parsing

mrshenli committed 3 years ago

cebe87fe

avoid inlining kernel lambdas on mobile (#46249)

smessmer committed 3 years ago

83286303

Back out "Revert D24269034: [fx] Refactor Tracer so that find_module and root args creation could be overridden by implementations" (#46573)

mrzzd committed 3 years ago

8357e2ed

[py][vulkan] Add is_vulkan to py api, add vulkan to device type parsing (#46511)

IvanKobzarev committed 3 years ago

e8fbe54c

preserve non-dense or overlapping tensor's layout in *_like functions (#46046)

lixinyu committed 3 years ago

a651b876

Revert D24004795: [quant] Add FixedQParamsFakeQuantize module

Ashkan Aliabadi committed 3 years ago

21814490

[caffe2][dnnlowp] Remove openmp usage in quantize dnnlowp op

dskhudia committed 3 years ago

f47231bf

Provide CMake option to enable Vulkan API. (#46503)

Ashkan Aliabadi committed 3 years ago

6cd8b5e9

Add Vulkan job dispatch and flush. (#46008)

Ashkan Aliabadi committed 3 years ago

3e041b50

Promote -Wcast-function-type to an error in builds. (#46356)

pritamdamania committed 3 years ago

cb3c1d17

Implement all communication APIs in DistributedC10d new frontend (#46053)

gmagogsfm committed 3 years ago

42a70dc5

[quant] Add FixedQParamsFakeQuantize module (#45538)

jerryzh168 committed 3 years ago

253918ec

[JIT] adding torch.jit.isinstance support (#46062)

Lilyjjo committed 3 years ago

f83cf2da

Support %-based string formatting (#45976)

Ansley Ussery committed 3 years ago

fdc5261a

[quant][refactor] Remove register api and rename get_*_mapping to get_default_*_mapping (#46337)

jerryzh168 committed 3 years ago

f9446cb1

Revert D24395956: [pytorch][PR] Replace flatten tensors with flatten loops.

Ashkan Aliabadi committed 3 years ago

4f5b55f7

Remove PyCFunction casts as much as possible. (#46227)

pritamdamania committed 3 years ago

2b221a95

[StaticRuntime] Threading model (#46219)

Hao Lu committed 3 years ago

1a3ea46d

Newer Older