strisunshinewentingwang/pytorch

Pull Requests Commits

Update TensorPipe submodule (#54507)

lw committed 3 years ago

35186eb9

Fix Flake8 (#54540)

samestep committed 3 years ago

1b792a7f

[ROCm] allow PYTORCH_ROCM_ARCH in cpp_extension.py (#54341)

jeffdaily committed 3 years ago

e5b97777

[complex] torch.rsub(): complex autograd support (#53702)

kshitij12345 committed 3 years ago

446e477d

gdb special command to print tensors (#54339)

antocuni committed 3 years ago

21a9a93e

[Pytorch Mobile] optimize_for_mobile: Fuse Add Relu on any function (#54441)

JacobSzwejbka committed 3 years ago

583c4bf7

disable cu112 test on windows (#54512)

Rong Rong (AI Infra) committed 3 years ago

acffa604

Reset Optimizer counter while deserializing netWithBackwardOptions

Neha Shah committed 3 years ago

f3c00047

Fix minor whitespace typo in tools/test_history.py (#54504)

samestep committed 3 years ago

ba9f12d2

Automated submodule update: FBGEMM (#54486)

facebook-github-bot committed 3 years ago

a4a21e7d

[jit][tensorexpr] Added aten::batch_norm into fuser when in inference mode (#54204)

huiguoo committed 3 years ago

2a538971

[AutoAccept][Codemod][FBSourceClangFormatLinter] Daily `arc lint --take CLANGFORMAT`

generatedunixname89002005325676 committed 3 years ago

fee470d8

Enabled BFloat16 support for argmax & argmin on both CPU & CUDA (#52582)

sanchit committed 3 years ago

f2a38a0e

Fix onnx warning message (#54371)

Peiyuan Liao committed 3 years ago

3519625a

Grammatically update tech docs (#54370)

Ioana Tivadar committed 3 years ago

1041fdd0

[PyTorch] Update Bazel build for TensorPipe (#54416)

lw committed 3 years ago

8518b0ee

[Gradient Compression] Remove cuda.syncrhonize in batched powerSGD (#54482)

Yi Wang committed 3 years ago

c22fc448

Delete all unnecessary singular Math entries (#54436)

ezyang committed 3 years ago

6d002719

s/StructuredNativeFunctions/NativeFunctionsGroup/ (#54427)

ezyang committed 3 years ago

6e8c4ad7

Rejigger to use NativeFunctionsGroup even without structured: True (#54426)

ezyang committed 3 years ago

bf2ca35f

Move compute_native_function_declaration to its own dest module (#54419)

ezyang committed 3 years ago

c00d66f7

Replace some tensor.device().is_cpu() calls with direct tensor.is_cpu() (#54397)

ezyang committed 3 years ago

349a17f1

[5/n][torch/elastic][upstream] Move torchelastic/agent to torch/distributed/elastic/agent (#54343)

aivanou committed 3 years ago

77ccd4f9

Port index_copy from TH to ATen (#52203)

lezcano committed 3 years ago

58703461

[Static Runtime] Fix bug in reshape_copy (#54467)

Hao Lu committed 3 years ago

52abd3bd

Only allow hub.load() from original repo. (#54451)

Ailing Zhang committed 3 years ago

c411017a

[JIT] Add Reinplacing to MKLDNN Subgraphs (#53908)

Elias Ellison committed 3 years ago

9be4c75f

use reshape when possible in broadcasting (#53326)

Elias Ellison committed 3 years ago

81c6e5fb

Avoid dispatch overhead in call to mkldnn convolution (#52614)

Elias Ellison committed 3 years ago

18c04a3f

[PyTorch][JIT] Less shared_ptr use in dictConstruct (#54110)

swolchok committed 3 years ago

3959d393

Newer Older