strisunshinewentingwang/pytorch

Pull Requests Commits

Symintify decomps for split and upsample_bilinear; Fix decomp for _softmax_backward_data and native_dropout_backward (#88761)

SherlockNoMad committed 2 years ago

9eabcc37

[primTorch] Add ref for `complex` (#88562)

nkaretnikov committed 2 years ago

76af7144

Skip dynamo benchmark tests under TSAN (#88895)

jansel committed 2 years ago

8f7e519f

meta function for max_pool2d_with_indices_backward (#88743)

anjali411 committed 2 years ago

52be0c42

Revert "[reland][dynamo] Better support for nn.Module (#88959)"

pytorchmergebot committed 2 years ago

98bcb4ac

[reland][dynamo] fixes dict changed during runtime error (#88877)

anijain2305 committed 2 years ago

897d029a

[Dynamo][FSDP] Migrate to `ModuleWrapPolicy` (#88453)

awgu committed 2 years ago

4284862d

Move xnnpack taget to fb code base (#88909)

cccclai committed 2 years ago

bca75fd2

[dynamo] Skip frame when graph break in a loop (#88857)

anijain2305 committed 2 years ago

2b12bfce

[reland][dynamo] Better support for nn.Module (#88959)

anijain2305 committed 2 years ago

e950afc3

[dynamo] Port all pytorch/dynamo and test/dynamo pieces over from symbolic-shapes branch (#88768)

voznesenskym committed 2 years ago

06ce1338

[FSDP] Fix `FSDP.clip_grad_norm_()` for `NO_SHARD` (#88955)

awgu committed 2 years ago

4f2639e5

Fix XLA symbolic shapes binding (#88928)

ezyang committed 2 years ago

46796fe5

Vectorized CPU code implementing left shift operator. (#88607)

alexsamardzic committed 2 years ago

2aca97cc

[16/N] Add _allgather_base custom op with CPU/CUDA implementation (#88889)

H-Huang committed 2 years ago

df1df9d1

torchdynamo support self.modules() for nn_module (#88695)

ydwu4 committed 2 years ago

37656213

Turn internal assert when saved tensor is detached inplace into torch check (#88860)

soulitzer committed 2 years ago

27dc03e0

[primTorch] Improve `narrow` and `narrow_copy`: refs, tests, docs (#87045)

nkaretnikov committed 2 years ago

4270bb37

[15/N] Add allreduce_coalesced custom op with CPU/CUDA implementations (#88846)

H-Huang committed 2 years ago

6e5f736d

Revert "[dynamo][api] Better support of torch.nn.Module (#88629)"

pytorchmergebot committed 2 years ago

ae2c668c

[quant][executorch] Support quant fusion for reshape in quant in executorch stack (#88858)

jerryzh168 committed 2 years ago

6b775c42

Revert "Add comprehensive minifier tests (#88022)"

pytorchmergebot committed 2 years ago

34641c43

[dynamo][api] Better support of torch.nn.Module (#88629)

anijain2305 committed 2 years ago

c8334859

[FSDP] Introduce `ModuleWrapPolicy` for simplicity (#88450)

awgu committed 2 years ago

d01bf1d1

[vision hash update] update the pinned vision hash (#88920)

pytorchmergebot committed 2 years ago

b2b0a0d3

[FSDP][state_dict][6/N] Remove most FSDP module dependency from _optim_utils (#88638)

fegin committed 2 years ago

ae407466

Exclude poolformer_m36 from the inductor model test (#88908)

desertfire committed 2 years ago

41083671

fix fx tests (#88886)

mikey dagitses committed 2 years ago

1e2327ba

Fix bug in OptionalTensorList (#88887)

ezyang committed 2 years ago

66736ff4

Remove incorrect assert about hermetic state. (#88885)

ezyang committed 2 years ago

2b166532

Newer Older