strisunshinewentingwang/pytorch

Pull Requests Commits

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

Remove padding_mode from torch.nn.functional.conv{1,2,3}d's docstr (#20891)

yifuwang committed 5 years ago

5e69e76a

Update nccl submodule to v2.4.6 (#20882)

mrshenli committed 5 years ago

4c5b1e34

Use a simpler way to delete recursive function

Stonesjtu committed 5 years ago

9310e600

fixed issue #20921 (#20922)

shagunsodhani committed 5 years ago

66e6571e

Update multiprocessing note now that shared CUDA tensors are refcounted (#19904)

ssnl committed 5 years ago

83fe9287

Fix pytorch_macos_10_13_py3_test (#20944)

ezyang committed 5 years ago

bdce5533

fix bug of not using get_score_cls_index in BoxWithNMSLimitOp (#20868)

wat3rBro committed 5 years ago

81e70ffa

Add warning about memory overhead when using multiple tiny tensors (#20801)

iamhatesz committed 5 years ago

2fb665a9

allow pass ordered dict for nn sequential (#20796)

wanchaol committed 5 years ago

c7e07228

Fix advanced indexing on "huge" Tensors (#20919)

colesbury committed 5 years ago

b93bdf69

Newer Older