strisunshinewentingwang/pytorch

Pull Requests Commits

masked_scatter thrust->cub (#56750)

zasdfgbnm committed 3 years ago

5152cf86

[DataLoader] Introduce MapMapDataPipe functional datapipe (#58258)

MarcioPorto committed 3 years ago

4942fe02

[DDP] Support not all outputs used in loss calculation (#57081)

rohan-varma committed 3 years ago

faa7d379

Fix dtype inference in sparse_csr_tensor_ctor (#58631)

malfet committed 3 years ago

abb215e2

Fix bug in test_fx_experimental codegen (#58587)

heitorschueroff committed 3 years ago

9ac0bd23

Enables builds with Compute Library backend for oneDNN (#55913)

nSircombe committed 3 years ago

bf00d26d

DOC Adds code comment to clarify nn.Linear.reset_parameters (#58487)

thomasjpfan committed 3 years ago

145a6f79

[pkg] Catch exceptions where dependency resolution gets invalid imports (#58573)

suo committed 3 years ago

5caccbe3

[pkg] simplifications to broken dependency handling (#58572)

suo committed 3 years ago

703f2439

Quote in setup-ci-env (#58637)

malfet committed 3 years ago

c4f0c5ee

Fix GIL issue when acquiring multiple sessions. (#58584)

Ailing Zhang committed 3 years ago

8615fd65

Make torch::deploy work with or without cuda (#58493)

wconstab committed 3 years ago

24786bd6

port `sgn` to structured (#58197)

eqy committed 3 years ago

fbc235c2

Port fmax & fmin to structured kernel (#58458)

Kiyosora committed 3 years ago

b5e39bce

[FX Splitter] dump final graph and print operator stats via to_glow API

Luoshang Pan committed 3 years ago

e179a568

refactor ASGD to use functional API (#58410)

iramazanli committed 3 years ago

9a622f4c

remove redundant getDispatchKeySetUnboxed(eligibleKeys) (#58535)

Basil Hosmer committed 3 years ago

208b36f1

Rename namespace `vec256` to `vec`, struct `Vec256` to `Vectorized` (and other related classes/structs) (#58438)

imaginary-person committed 3 years ago

47c566eb

Revert D28461013: [nnc] Enable CPU fusion inside Facebook, take 2

bertmaher committed 3 years ago

a6b358d5

[FX] Add APIs to mutate specific args/kwargs (#58571)

James Reed committed 3 years ago

36adc3f0

Update on "[c10d] Introduce ProcessGroupWrapper"

rohan-varma committed 3 years ago

677d3705

Update base for Update on "[c10d] Introduce ProcessGroupWrapper"

rohan-varma committed 3 years ago

151caaa0

[THC] Rename THCTensorMathMagma from cu to cpp (#58521)

malfet committed 3 years ago

296d2a43

Added publishing of test results and minor fixes to Az DevOps Build Logic (#58436)

mstfbl committed 3 years ago

ae99640a

[profile] fix recorded data type (#58531)

kingchc committed 3 years ago

b9b8522e

Revert "Move Azure MultiGPU tests back to nightly (#58242)" (#58451)

skyline75489 committed 3 years ago

8de8b492

Fix some tensor operators to return `NotImplemented` for invalid inputs (#58216)

asi1024 committed 3 years ago

3113a1de

step 0 of cuDNN v8 convolution API integration (#51390)

zasdfgbnm committed 3 years ago

6c70cbed

[ATen][Quant] Pass at::Tensor by reference (#58284)

Hao Lu committed 3 years ago

954d39ba

model_dump: Accept variable-length debug info (#57660)

dreiss committed 3 years ago

a9137543

Newer Older