strisunshinewentingwang/pytorch

Pull Requests Commits

Add a shortcut to test all torchbench models. (#57311)

xuzhao9 committed 3 years ago

d68ad3cb

torch.clamp with tensor min and max (#52695)

peterbell10 committed 3 years ago

33eea146

Port trunc to structured (#57350)

Freey0 committed 3 years ago

c328bb6d

[16/n] [torch/elastic] Introduce _RendezvousOpExecutor (#57144)

cbalioglu committed 3 years ago

1a6f827a

[15/n] [torch/elastic] Introduce _RendezvousStateHolder (#56538)

cbalioglu committed 3 years ago

76bccfb2

fix comments in ATenNVRTC.h (#57318)

eqy committed 3 years ago

160304a8

[RELAND] [CUDA graphs] Avoid sync errors when graph capturing cudnn rnn calls that use cudnn dropout (#57373)

mcarilli committed 3 years ago

e841f335

[14/n] Introduce a name attribute to _PeriodicTimer (#57143)

cbalioglu committed 3 years ago

1b745efb

[13/n] Extend the return type of RendezvousBackend's set_state method (#57142)

cbalioglu committed 3 years ago

233004b4

[12/n] Rename last_keep_alives to last_heartbeats in _RendezvousState (#57141)

cbalioglu committed 3 years ago

a6f60cf4

[11/n] [torch/elastic] Add heartbeat timeout to RendezvousTimeout (#57140)

cbalioglu committed 3 years ago

32093647

[10/n] [torch/elastic] Add comparison operators to _NodeDesc (#57139)

cbalioglu committed 3 years ago

6876e15d

[9/n] [torch/elastic] Introduce RendezvousSettings (#56537)

cbalioglu committed 3 years ago

6bf8df6b

[PyTorch][Edge] Add api to get bytecode version from runtime (#56948)

cccclai committed 3 years ago

ac71432c

[quant][graphmode][fx] Skip observering boolean Tensors (#57375)

jerryzh168 committed 3 years ago

945c93b8

Use ld.gold by default to link in CI (#57061)

davidriazati@fb.com committed 3 years ago

264d8798

Replace 11.2 linux CI with 11.3 (#57222)

janeyx99 committed 3 years ago

c0d39ba6

[DDP] Profile search_unused_parameters (#57376)

rohan-varma committed 3 years ago

375c8a81

Port max_pool2d_with_indices to structured kernel (#56459)

liuyuanqiang@bytedance committed 3 years ago

52b38925

Revert D28143091: [pytorch][PR] Add cross OpInfo

mrshenli committed 3 years ago

6bc3ad28

[BE] Improve has_bf16_support (#57408)

malfet committed 3 years ago

c7d8d8f9

Implement result() function in MPI Work classes (#57168)

agolynski committed 3 years ago

f332a8bd

use importlib instead of imp as it support python 3.5+ (#57160)

albanD committed 3 years ago

0a0e0246

Automated submodule update: tensorpipe (#56916)

facebook-github-bot committed 3 years ago

7e12c3e1

Eliminate global usage of torch.set_default_dtype in test_autograd (#56446)

aocsa committed 3 years ago

87242d23

OpInfo: ravel, view, view_as (#56910)

kshitij12345 committed 3 years ago

154eca03

Assert that GIL is not held in blocking destructors (#57030)

ezyang committed 3 years ago

e845158b

fx quant: remove `find_quants` from convert (#57402)

vkuzo committed 3 years ago

da51fd31

fx quant: remove unnecessary quants arguments (#57399)

vkuzo committed 3 years ago

d6563bc1

fx quant: remove FixedQParamsOpQuantizeHandler from quantize.py (#57393)

vkuzo committed 3 years ago

643f41be

Older