strisunshinewentingwang/pytorch

Pull Requests Commits

sparse.mm backward: performance improvements (#94991)

nikitaved committed 1 year ago

056d92e2

Inductor Freezing (#100652)

eellison committed 1 year ago

d083d444

CUDA graphs overrides dynamic shapes and forces specialization (#103290)

ezyang committed 1 year ago

54daf870

Add graph break logging option instead of config flag (#103202)

mlazos committed 1 year ago

6c6c897d

[c10d] Add xpu to the default device supported by user specified backend (#103410)

zhuhong61 committed 1 year ago

50c972bf

Rewrite size/stride/numel TensorVariable handling (#103438)

ezyang committed 1 year ago

49754f44

[CI] Update inference accuracy test (#103361)

desertfire committed 1 year ago

14182849

Update C++ frontend docs (#103451)

malfet committed 1 year ago

f22d99c7

[Reland] Add sym_size/stride/numel/storage_offset to native_function.yaml (#103107)

SherlockNoMad committed 1 year ago

d997969b

Revert "Move tensor grouping to ATen (#100007)"

pytorchmergebot committed 1 year ago

0cb5bc3b

Add uint8 support for CPU images in interpolate(mode='bicubic') (#103252)

NicolasHug committed 1 year ago

3766c047

Revert "change pre_autograd to pre_dispatch tracing (#101818)"

pytorchmergebot committed 1 year ago

5ed61813

Make all CI commit pin changes trigger ciflow/inductor (#103443)

ezyang committed 1 year ago

1b31665e

Revert "Cleanup scatter-related code (#103074)"

pytorchmergebot committed 1 year ago

fc46f01b

[ONNX] Fix onnx constant folding (#101329)

YJ Shi committed 1 year ago

1ef7d679

enable missing-prototypes in functorch (#103391)

cyyever committed 1 year ago

48e3ee29

Replace CUDA 11.7 small pip wheels with 12.1 (#103091)

ptrblck committed 1 year ago

de354bf5

Remove dynamic_shapes special case in SizeVariable getitem (#103380)

ezyang committed 1 year ago

ac3ce0a5

[dynamo][numpy] Support ndarray methods (#97537)

larryliu0820 committed 1 year ago

2eac8bd2

Clean up op BC check list (#103363)

houseroad committed 1 year ago

18f203a5

[dynamo] graph break on nn.Parameter construction (#103262)

davidberard98 committed 1 year ago

df83fe5b

Revert "Update torchbench pin - torchrec_dlrm moved to canary (#103383)"

pytorchmergebot committed 1 year ago

08f90b34

Revert "Log functional_collectives apis to distributed logger (#103288)"

pytorchmergebot committed 1 year ago

caecb552

Always create ShapeEnv, always apply unspec logic (#103302)

ezyang committed 1 year ago

c3fdfca5

[xla hash update] update the pinned xla hash (#103416)

pytorchupdatebot committed 1 year ago

f4228e70

Log functional_collectives apis to distributed logger (#103288)

wconstab committed 1 year ago

37359c36

[Quant][XNNPACK] Delegate add_relu fusion (#103266)

mcr229 committed 1 year ago

f37be778

Up to 48% speed up using Sklansky method for innermost prefix scan algorithm (#103314)

mfkasim1 committed 1 year ago

8a744c31

Handle nonzero via its meta registration (#103379)

ezyang committed 1 year ago

0863e550

Update torchbench pin - torchrec_dlrm moved to canary (#103383)

anijain2305 committed 1 year ago

114f99bb

Newer Older