strisunshinewentingwang/pytorch

Pull Requests Commits

[quant][ao_migration] `torch.nn.quantizable` → `torch.ao.nn.quantizable`. (#78717)

zaf committed 2 years ago

29e83b65

[quant][ao_migration] `torch.nn.quantized._reference` → `torch.ao.nn.quantized._reference` (#78715)

zaf committed 2 years ago

b1455f94

[quant][ao_migration] `torch.nn.quantized.dynamic` → `torch.ao.nn.quantized.dynamic` (#78714)

zaf committed 2 years ago

d32a7621

[quant][ao_migration] `torch.nn.quantized.modules` → `torch.ao.nn.quantized.modules` (#78713)

zaf committed 2 years ago

c92e5ac9

enable qlinear dynamic parallelization with fbgemm (#84033)

jianyuh committed 2 years ago

a83d7d8b

Decomposition - batch_norm, save_mean and save_variance always float32 (#84013)

anijain2305 committed 2 years ago

e2f75d63

fix `NoneType` object has no attribute `python_exit_status` (#83985)

ejguan committed 2 years ago

56fef4e6

[functorch] add batching rule for fill_.Tensor (#84015)

zou3519 committed 2 years ago

00cb1845

add qscheme check for quantization observer (#80126)

XiaobingSuper committed 2 years ago

31f15176

Make linalg.inv composite of linalg.solve (#80074)

lezcano committed 2 years ago

f5a35150

Disable autocast cache during aotdispatch (#84035)

Chillee committed 2 years ago

e3c89d07

switching the exact check to isinstance check (#84023)

Krovatkin committed 2 years ago

63cbdc92

Enable cache action for lint workflow (#84026)

huydhn committed 2 years ago

02c37813

[fx] add deferred weights (xl_weight) and tracing for xl_embedding_bag (#84016)

alexbeloi committed 2 years ago

c00f0c80

Fix dumb make_fx issue (#84011)

Chillee committed 2 years ago

8b8942b1

[fx+scripting] Adding num_iter_1 and num_iter_2 params LearningRate op (#83691)

Mandar Deshpande committed 2 years ago

c03f8abb

add test comparing fast and slow path of transformer encoder. fix src_key_padding_mask bug with seqlen > 1024

erichan1 committed 2 years ago

2186337f

erichan1 committed 2 years ago

d39c6497

update test to catch seqlen 1024 fastpath bug

erichan1 committed 2 years ago

bcfbcae3

NCCL: Re-enable parallel builds (#83696)

peterbell10 committed 2 years ago

2000eba4

Revert "[TorchTidy] Adding support for unique tensor identifiers (#80266)"

pytorchmergebot committed 2 years ago

d5af2a70

Revert "Support NCCL Premul Sum (#81272)"

pytorchmergebot committed 2 years ago

1f61c39a

[caffe2] Remove last clang-for-cuda sources (#84021)

andrewjcg committed 2 years ago

460636ab

fix oneDNN channels_last path issue (#83653)

XiaobingSuper committed 2 years ago

a013597b

[TorchTidy] Adding support for unique tensor identifiers (#80266)

Gamrix committed 2 years ago

b6ba4192

[NVFuser] Upstream push 0811 (#83239)

jjsjann123 committed 2 years ago

b21a6ff6

Increase timeout for linux binary builds (#84008)

atalman committed 2 years ago

e90db175

[Quant] Vectorize scalar remainder in quantized kernel for normalization (#79673)

Xia-Weiwen committed 2 years ago

6b597595

Revert "Don't introduce new overload for SymInt (#83628)"

pytorchmergebot committed 2 years ago

a7edf713

Revert "[xla hash update] update the pinned xla hash (#83967)"

pytorchmergebot committed 2 years ago

7a02ee55

Newer Older