strisunshinewentingwang/pytorch

Pull Requests Commits

Clarify the saving of intermediates in the "extending torch.func" docs (#98020)

zou3519 committed 1 year ago

ebe08e11

[inductor] use triu ref instead of lowering (#96040) (#96462)

atalman committed 1 year ago

Verified c263bd43

Add jinja2 as mandatory dependency (#95691) (#96450)

atalman committed 1 year ago

Verified c9913cf6

Fix expired deprecation of comparison dtype for NumPy 1.24+ (#91517) (#96452)

atalman committed 1 year ago

Verified 2f7d8bbf

`dl_open_guard` should restore flag even after exception (#96231) (#96457)

malfet committed 1 year ago

Verified ca0cdf52

[Release/2.0] Use Triton from PYPI (#96010)

malfet committed 1 year ago

Verified 9cfa076d

[Release/2.0] Use builder release branch for tests

malfet committed 1 year ago

8e05e41d

Remove mention of dynamo.optimize() in docs (#95802) (#96007)

msaroufim committed 1 year ago

Verified d8ffc60b

[MPS] Disallow reshape in slice (#95905) (#95978)

kulinseth committed 1 year ago

Verified 14837230

[MPS] Add fixes for div with floor (#95869)

kulinseth committed 1 year ago

Verified c4572aa1

[MPS] Fix views with 3 or more sliced dimensions (#95762) (#95871)

kulinseth committed 1 year ago

Verified 82b078ba

Remove torch._inductor.config.triton.convolution (#95840)

jansel committed 1 year ago

Verified 77f7bc5f

[optim] _actually_ default to foreach (#95862)

janeyx99 committed 1 year ago

Verified 08659645

Release version of fixed nll decomp (#95853)

ngimel committed 1 year ago

Verified f18ac1b3

[ROCM] Restrict pytorch rocm to only use triton 2.0.x (#95793) (#95834)

dllehr-amd committed 1 year ago

Verified c04134cd

[BE] Fix TORCH_WARN_ONCE (#95559) (#95822)

malfet committed 1 year ago

Verified 72d0863a

Update copyright (#95652) (#95700)

atalman committed 1 year ago

Verified 1bd334dc

[MPS] Remove FFT from the fallback as its causing crashes in test_ops and TestConsistency tests. (#95625)

kulinseth committed 1 year ago

Verified 93e13cd4

[MPS] Add TORCH_CHECK for Convolution (#95495)

kulinseth committed 1 year ago

Verified 4e4d4b0a

Reserve the tensorrt backend name for torch-tensorrt (#95627)

Wei committed 1 year ago

Verified c4fa8508

Add float to list of allowed ops (#94910) (#95661)

atalman committed 1 year ago

Verified 36ead098

fix spurious aot autograd warning (#95521) (#95614)

bdhirsh committed 1 year ago

Verified 66d23dba

[CUDA][CUBLAS] Explicitly link against `cuBLASLt` (#95094) (#95615)

atalman committed 1 year ago

Verified e2fff588

Update triton hash (#95540) (#95577)

weiwangmeta committed 1 year ago

Verified 735333a7

[MPS] LSTM fixes (#95388)

kulinseth committed 1 year ago

Verified 60174888

[optim] Add general documentation on our algorithm defaults (#95391) (#95516)

janeyx99 committed 1 year ago

Verified e51e5e72

hotfix for memory leak in aot autograd induced by saving tensors for backward (#95101) (#95477)

bdhirsh committed 1 year ago

Verified 91739a02

inductor: fix complier error when trying to vectorize logit_and and logit_or (#95361) (#95439)

XiaobingSuper committed 1 year ago

Verified 531f097b

[optim] Set defaults to foreach, NOT fused (#95241) (#95415)

janeyx99 committed 1 year ago

Verified 00eb7b0d

[SDPA] Fix bug in parsing scaled_dot_product_attention arguments (#95311) (#95397)

drisspg committed 1 year ago

Verified 2180f342

Older