strisunshinewentingwang/pytorch

Pull Requests Commits

[Quant] Add initial Executorch BackendConfig

andrewor14 committed 3 years ago

cd9372ad

Make CUDA exceptions unlikely and isolate C10_CUDA_CHECK body (#85256)

r-barnes committed 3 years ago

f0869cc8

Revert "[Profiler] Make `LibKinetoClient::stop()` directly call `ProfilerStateBase::pop` (#83965)"

pytorchmergebot committed 3 years ago

f0a084f3

Skip MPS test from generic M1 testsuite (#85500)

malfet committed 3 years ago

46a6a50f

disable circleci jobs b/c they are flaky (#85508)

Catherine Lee committed 3 years ago

92a94210

Use fallback approach for nested matmul (#85311)

mikaylagawarecki committed 3 years ago

5e700803

[Static Runtime] Fold linear prepack ops (#85289)

Mike Iovine committed 3 years ago

63c1f2fe

[Static Runtime] Fix aten::index_put list conversions (#85298)

Mike Iovine committed 3 years ago

e4899764

Increase timeout for ProcessGroupGlooTest (#85474)

Flamefire committed 3 years ago

bd854588

Revert "[mta] APEX style Fused Adam (#81705)"

pytorchmergebot committed 3 years ago

e505360e

Delete functorch's monkeypatching (#85430)

zou3519 committed 3 years ago

84843759

Move functorch python bindings to torch/csrc (#85426)

zou3519 committed 3 years ago

5e5c3195

Remove deprecated torch.matrix_rank (#70981)

IvanYashchuk committed 3 years ago

bcf93181

Adding conda retry upload to mitigate connection reset errors (#85407)

atalman committed 3 years ago

e3429769

Exposing native _scaled_dot_product_attention to torch.nn (#85044)

drisspg committed 3 years ago

253ffbf2

[GHF] Create "Core Reviewers" group (#85477)

malfet committed 3 years ago

0d04e548

Disabling the pypi cudnn wheel from uploading temporarily (#85470)

atalman committed 3 years ago

e227e3ec

Revert "Add FakeCrossRef tests for backwards, Fix Layer Norm Backward Decomp (#85417)"

pytorchmergebot committed 3 years ago

5043457a

Apply new symbolic shape strategy to make_fx symbolic mode (#85260)

ezyang committed 3 years ago

9baf6770

[ONNX] Update docs on symbolic registration (#85290)

justinchuby committed 3 years ago

2f50d2f6

Add FakeCrossRef tests for backwards, Fix Layer Norm Backward Decomp (#85417)

eellison committed 3 years ago

9c770839

More SymFloat support (#85411)

ezyang committed 3 years ago

61b4e8a7

[maskedtensor] add basic tests and unary/binary/reduction tests from common_method_invocations (#82841)

george-qi committed 3 years ago

0c46e3ec

Reduce memory usage requirement of test_warp_softmax_64bit_indexing in test_nn.py (re-open of #85037) (#85373)

eqy committed 3 years ago

2bc82163

[ONNX] Use decorators for symbolic function registration (#84448)

justinchuby committed 3 years ago

76d60778

Skip NVIDIA driver installation if it's already there (#85435)

huydhn committed 3 years ago

c7c2578f

[vision hash update] update the pinned vision hash (#85451)

pytorchmergebot committed 3 years ago

99ad8a30

SubgraphMatcher remove invalid matches (#85444)

SherlockNoMad committed 3 years ago

34296e2f

[quant][docs][ez] Fix formatting for qconfig_mapping (#85306)

jerryzh168 committed 3 years ago

4523ac7a

[mkldnn_matmul] enable mkldnn matmul for aarch64 bf16 devices (#83671)

snadampal committed 3 years ago

f21e77d9

Older