strisunshinewentingwang/pytorch

Pull Requests Commits

Instantiate PackedConvWeight to avoid linking error (#49442)

iseeyuan committed 3 years ago

f2ee8c62

CUDA BFloat embedding (#44848)

zasdfgbnm committed 3 years ago

86902f84

webdataset prototype - LoadFilesFromDiskIterableDataset (#48955)

lixinyu committed 3 years ago

001ff3ac

webdataset prototype - ListDirFilesIterableDataset (#48944)

lixinyu committed 3 years ago

6786b2b9

Enhanced generators with grad-mode decorators (#49017)

ivannz committed 3 years ago

efc09065

[PyTorch] Avoid move-constructing a List in listConstruct (#49355)

swolchok committed 3 years ago

76d09ec3

Making ops c10-full: optional lists (#49088)

smessmer committed 3 years ago

ec8e9d31

Making ops c10 full: optional out arguments (#49083)

smessmer committed 3 years ago

d69d42db

Revert D25554109: [StaticRuntime][ATen] Add out variant for narrow_copy

Hao Lu committed 3 years ago

306bab22

[StaticRuntime][ATen] Add out variant for narrow_copy (#49449)

Hao Lu committed 3 years ago

ed04b716

Unescape string in RPC error message (#49373)

rohan-varma committed 3 years ago

40d7c109

quantized tensor: add preliminary support for advanced indexing, try 2 (#49346)

vkuzo committed 3 years ago

a9137aeb

[StaticRuntime] Fusion pass for ClipRanges/GatherRanges/LengthsToOffsets (#49113)

Hao Lu committed 3 years ago

8954eb3f

fix optimizer.pyi typo 'statue'->'state' (#49388)

lixinyu committed 3 years ago

94e328c0

[StaticRuntime] Permute_out (#49447)

Hao Lu committed 3 years ago

cbeb4c25

update breathe (#49407)

mattip committed 3 years ago

acd72e79

[CMake] Use libtorch_cuda list defined in bzl file (#49429)

malfet committed 3 years ago

58551e52

[PyTorch] Use plain old function pointer for RecordFunctionCallback (reapply) (#49408)

swolchok committed 3 years ago

22c6dafd

[FX] Rename Node._uses and refactor Node.all_input_nodes (#49415)

James Reed committed 3 years ago

e9d7d37a

[DPER] Introduce barrier operation to force synchronization of threads in async execution (#49322)

kennyhorror committed 3 years ago

46debe7f

Add flag torch_jit_disable_warning_prints to allow disabling all warnings.warn (#49313)

gmagogsfm committed 3 years ago

7518f546

Fix include files for out-of-tree compilation (#48827)

CaoZhongZ committed 3 years ago

aff0b68a

Replace THError() check in THCTensorMathReduce.cu with C10_CUDA_KERNEL_LAUNCH_CHECK() (#49424)

Amogh Akshintala committed 3 years ago

16f4b0ed

[numpy] torch.{all/any} : output dtype is always bool (#47878)

kshitij12345 committed 3 years ago

c508e5b1

[JIT] Support multiple outputs in subgraph matcher. (#48992)

Mikhail Zolotukhin committed 3 years ago

38a59a67

[static runtime] refine fusion group (#49340)

bwasti committed 3 years ago

3ffe9e0f

[te] Fix bugs with shift operators (#49396)

bertmaher committed 3 years ago

f4e15c4a

Making ops c10-full: Generator arguments (#49013)

smessmer committed 3 years ago

5912316c

Making ops c10 full: out overloads with default arguments (#49012)

smessmer committed 3 years ago

a6274c12

Making ops c10-full: Dimname arguments (#49008)

smessmer committed 3 years ago

b47fa5e8

Newer Older