strisunshinewentingwang/pytorch

Pull Requests Commits

Report timeout/infra_error instead of 0.0000 on infra error (#100372)

ezyang committed 1 year ago

5d93265c

[Easy][FSDP] Clarify `_use_unsharded_grad_views` comment (#100359)

awgu committed 1 year ago

a014d1b1

Refactor repro/minifier into CLI; add analyze (#100226)

ezyang committed 1 year ago

2d8deffc

Revert "Produce constant variables in cases where a SymNode is created with a constant (#100144)"

pytorchmergebot committed 1 year ago

89c43f41

[FSDP] Fix `use_orig_params=True`, CPU offload, `no_sync()` (#100180)

awgu committed 1 year ago

83b803c2

[BE] Fix SIM109 `compare-with-tuple` (#100337)

justinchuby committed 1 year ago

e779a30d

[BE] Fix all B022 `useless-contextlib-suppress` (#100335)

justinchuby committed 1 year ago

01abbfba

Produce constant variables in cases where a SymNode is created with a constant (#100144)

voznesenskym committed 1 year ago

d7bdfd34

[inductor] avoid zero division error for dropout (#100222)

Valentine233 committed 1 year ago

cc3ed8ae

Fix intermediate hooks on inplace buffers, enable it in testing (#100322)

ezyang committed 1 year ago

beb7f795

Use sympy.And instead of bitwise operator, for better promotion (#100328)

ezyang committed 1 year ago

155fa4e7

Skip invalid grads in outplace foreachs' backward (#100256)

crcrpar committed 1 year ago

6c934a89

fix TIMM mobilevit_s complier issue for dynamic CPU path (#100230)

XiaobingSuper committed 1 year ago

76bcc872

[decomp] Bad accuracy for elu_backward (#100284)

anijain2305 committed 1 year ago

e1021ec5

Revert "Move win-vs2019 build and test to unstable (#100281)"

pytorchmergebot committed 1 year ago

3d55bce3

[MPS] Fix `layer_norm_backward_mps` key (#100295)

malfet committed 1 year ago

2442858f

[dynamo] Compile torchvision augmentations (#100292)

anijain2305 committed 1 year ago

03806edd

Update docstring for dynamo.export tracing_mode (#100205)

BowenBao committed 1 year ago

6647e61a

Revert "Run test_fx_to_onnx_with_onnxruntime serially (#100298)"

pytorchmergebot committed 1 year ago

9075e3c2

[ONNX] Skip flaky dynamic test in CI (#100297)

BowenBao committed 1 year ago

1267897d

Run test_fx_to_onnx_with_onnxruntime serially (#100298)

huydhn committed 1 year ago

3a3f781f

Add size check before calling .back() in rpc/script_call.cpp (#94297)

m4drat committed 1 year ago

7684044b

fix(docs): torch.autograd.graph.Node.register_hook can override grad_inputs, not grad_outputs (#100272)

olokshyn committed 1 year ago

35991df5

Update testing aggregate data (#100070)

PaliC committed 1 year ago

2b79d6c4

Check inputs have same dtype in addmm_impl_cpu_ even if input has zero numel (#100274)

soulitzer committed 1 year ago

6a023421

Introduce fast path in the CPU equal op

houseroad committed 1 year ago

Verified d7fa7fa8

Add comment link to revert message (#100276)

clee2000 committed 1 year ago

331ed5be

Move win-vs2019 build and test to unstable (#100281)

ZainRizvi committed 1 year ago

999e17d8

follow up PR for test_c10d_ucc.py in response to Xiang's review of #88110 (#99654)

Fuzzkatt committed 1 year ago

ccce7a2d

[ONNX] Set tracing_mode through options.dynamic_shapes and enable dynamic tests in test_fx_to_onnx_runtime.py (#100212)

titaiwangms committed 1 year ago

8714fc7a

Newer Older