strisunshinewentingwang/pytorch

Pull Requests Commits

Re-land "Fix advanced indexing on "huge" Tensors" (#21019)

colesbury committed 5 years ago

b85c5292

Improve error message for missing attribute (#20779)

Chillee committed 5 years ago

52d27890

Some name and variable cleanup (#20861)

orionr committed 5 years ago

bc10677f

Re-enable test_dag_net_forking on ROCm (#21013)

bddppq committed 5 years ago

99674eb8

Clarify cycliclr param docs (#20880)

sampepose committed 5 years ago

082936f0

Change bound shape inference for LengthsRangeFill & GatherRanges, add more tests (#20610)

ChunliF committed 5 years ago

68c3ef72

Refactor schema_matching.cpp (#20549)

davidriazati committed 5 years ago

bbe34118

Generate TH functions outside of Type (#20309)

Roy Li committed 5 years ago

ff6cda0d

Move 1d tensor checks to TH (#20859)

Roy Li committed 5 years ago

eacb3118

Change view dispatch to abstract (#20308)

Roy Li committed 5 years ago

d2f14db6

Restore TBB module (#20454)

Ilia Cherniavskii committed 5 years ago

580eab65

Native ATen/Parallel backend (#20087)

Ilia Cherniavskii committed 5 years ago

82aecfad

Fix incorrect torch version in CMake (#21007)

peterjc123 committed 5 years ago

f4b434a6

fix small typo muliprocessing -> multiprocessing (#20998)

jpgard committed 5 years ago

05561413

Revert D15501945: [pytorch][PR] Fix advanced indexing on "huge" Tensors

bddppq committed 5 years ago

5ddbfc97

Check for incompatible versions between CUDA and MSVC

peterjc123 committed 5 years ago

3b0d431b

Update cuSPARSE namespace collision w/ CUDA 10.1 Update 1 (#20889)

syed-ahmed committed 5 years ago

0d35f145

Convert dequantize_linear to an internal function _dequantize_linear (#20938)

nishantpdce committed 5 years ago

9d9751f6

Remove functionality unsupported by the JIT from multi_head_attention_forward. (#20653)

Guanheng Zhang committed 5 years ago

8e3311c5

fix SyncBatchNorm doc (#20991)

soumith committed 5 years ago

6e76813a

fix the bug for mkldnn clone (#20943)

XiaobingSuper committed 5 years ago

ebc8d717

Revert D15511921: [pytorch][PR] BatchSampler now uses list.clear() instead of creating new objects

soumith committed 5 years ago

6480d3f1

BatchSampler now uses list.clear() instead of creating new objects

ssnl committed 5 years ago

482ae8e6

Update submodule URL based on redirection. (#20973)

xkszltl committed 5 years ago

ecf01221

Update cuda pinned memory note to include tensor.to (#20977)

ssnl committed 5 years ago

bb89827e

In setup.py, also check some submodules of submodules. (#20937)

xuhdev committed 5 years ago

1e8f129a

tweak tqdm to have download speed in kB/MB/etc (#20908)

djsutherland committed 5 years ago

8dbdd00f

.view(...) now suggests .reshape(...) instead .contiguous().view(...)

ssnl committed 5 years ago

5ab6e071

Fix subscripts in RNN documentation

Alexander Krotov committed 5 years ago

c611630b

Fix align corner docs (#20961)

daquexian committed 5 years ago

a3a458ed

Newer Older