strisunshinewentingwang/pytorch

Pull Requests Commits

Use standard docker image for XLA build

kostmo committed 5 years ago

b312e9de

Modernize test_sparse. (#17324)

gchanan committed 5 years ago

25730f15

remove nn.Upsample deprecation warnings from tests (#17352)

soumith committed 5 years ago

c63af883

upgrade documentation in setup.py to NO_ -> USE_ (#17333)

soumith committed 5 years ago

3069c450

Enforce non-negativity of tensor construction (#17077)

Dmytro Dzhulgakov committed 5 years ago

5744d521

Fixing docstring in CTCLoss (#17307)

Igor Macedo Quintanilha committed 5 years ago

94a95a0c

Fix the slowness of mvn's log_prob (#17294)

fehiepsi committed 5 years ago

de81a274

Move argsort to C++

zasdfgbnm committed 5 years ago

722cbe30

Include vec256 headers in setup.py (#17220)

tridao committed 5 years ago

37890610

Enable MAX_JOBS for using Ninja on Windows

peterjc123 committed 5 years ago

51069186

Avoid unnecessary CPU-to-GPU copy of torch.load with CUDA (#17297)

lw committed 5 years ago

29f4f8f0

allow lists to contain any tensor type (#17321)

suo committed 5 years ago

2c302b6e

Skip convnets benchmark in rocm CI (#17331)

bddppq committed 5 years ago

d92ddcf7

Don't have malloc-free pairs that cross DLL boundaries. (#17302)

ezyang committed 5 years ago

b3b692a8

Add support to build for multiple amd gpu targets (#17329)

bddppq committed 5 years ago

c063a33e

batched cleanups (#17288)

suo committed 5 years ago

501d346d

(Permanently) fix CI breakage due to new docker version. (#17338)

ezyang committed 5 years ago

cb3d3d11

Implementation convolutionTranspose operator for mkl-dnn (#12866)

PenghuiCheng committed 5 years ago

376bb403

Support multi-device configuration for MKL-DNN (#12856)

PenghuiCheng committed 5 years ago

c02e2ff0

fix missing std (#17263)

Ailing Zhang committed 5 years ago

90950f79

Rethrow exceptions from RunAsync (#15034)

Ilia Cherniavskii committed 5 years ago

0edc8113

Reinforce scheduling invariants (#17132)

Ilia Cherniavskii committed 5 years ago

0337494c

Modify TileOp GPU implementation to expose more concurrency and better utilize GPU memory bandwidth (#17275)

wesolwsk committed 5 years ago

3e44880d

Support str for native_functions.yaml schema

cpuhrsch committed 5 years ago

9e4a9938

Separate gpu reduce functions (#17146)

xiaomengy committed 5 years ago

2e67b34e

Minor doc updates in c10/core/Allocator.h (#17164)

ezyang committed 5 years ago

474adf54

Namedtuple return for symeig, eig, pstrf, qr, geqrf (#16950)

zasdfgbnm committed 5 years ago

b2dde438

Allow PyTorch to be built without NCCL (#17295)

t-vi committed 5 years ago

36ddad3b

add foxi submodule (#17184)

houseroad committed 5 years ago

e6cf3c88

Removed obsolete argument correct_transform_coords in bbox_transform op. (#16723)

newstzpz committed 5 years ago

54e4c4d7

Newer Older