strisunshinewentingwang/pytorch

Pull Requests Commits

[quant][improvement][better-engineering] Refactored get_supported_device_types into common_quantization.py (#79607)

dzdang committed 1 year ago

ea81138b

Faster mul(sparse, sparse) with broadcasting in dense dims. (#85336)

nikitaved committed 1 year ago

12ae3bea

Temporary fix to skip NVIDIA driver installation from RHEL repo (#85569)

huydhn committed 1 year ago

40d3e55b

[FX] Add one option to maintain the FX graph execution order after splitting_module (#85188)

RenfeiChen-FB committed 1 year ago

4befe450

Add torch.nested to ovrsource (#85384)

George Gensure committed 1 year ago

4f5f2c1a

Symintifying slice ops (#85196)

ezyang committed 1 year ago

4c01c512

OpInfo for Slice (#85554)

ezyang committed 1 year ago

604487f2

[fix] composite compliance: cumprod, _masked.cumprod, linalg.vander (#85330)

kshitij12345 committed 1 year ago

bc6dc8d2

[Quant] Add initial Executorch BackendConfig (#85527)

andrewor14 committed 1 year ago

2e817103

[Checkpoint] rename apply_ac_wrapper (#85449)

rohan-varma committed 1 year ago

a8074a1a

[Docs] Minor fix to apply_ac doc (#85448)

rohan-varma committed 1 year ago

cc64f646

Fix cuda issue with sparse.sampled_addmm (#85194)

srossross committed 1 year ago

a4c94f07

[ci] test_ops in parallel, ci tests log to file (#85528)

clee2000 committed 1 year ago

49e10c15

[NVFuser] Upstream push 0907 (#84626)

jjsjann123 committed 1 year ago

0e582fbf

Adjust retry time for conda upload (#85545)

atalman committed 1 year ago

52a8be52

Add new cudnn buid for linux only (#85549)

atalman committed 1 year ago

30070930

[CI] Make `cuda-arch-list` a parameter to linux-build (#85523)

malfet committed 1 year ago

d83ca9eb

Let antoniojkim snoop all symbolic shapes PRs (#85555)

ezyang committed 1 year ago

108b25db

[Profiler] Clean up Tensor representation (#85161)

Taylor Robie committed 1 year ago

4dfaca6f

[Profiler] Capture storage data pointer (#84276)

Taylor Robie committed 1 year ago

e296a82f

resubmit: [mta] APEX style Fused Adam (#81705) (#85507)

crcrpar committed 1 year ago

4615d1bc

[DataLoader] Make distributed lazily initialized & share seed via PG (#85279)

ejguan committed 1 year ago

f1a6f32b

[maskedtensor] move __torch_function/dispatch__ functions to a map (#85529)

george-qi committed 1 year ago

e3766e98

Add instructions on how to merge a PR (#85280)

ZainRizvi committed 1 year ago

78937489

Add nvprims `rand_like` support for Dropout (#85077)

kevinstephano committed 1 year ago

c7b17d7e

Fake tensor refactorings (#85498)

eellison committed 1 year ago

1e4c8851

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

pytorchmergebot committed 1 year ago

d10de31c

Revert "add amp tests (#85434)"

pytorchmergebot committed 1 year ago

eb570ab7

Revert "Turn on aliasing tests for fake backwards, Fix Batch norm running mean/var decomp aliasing (#85471)"

pytorchmergebot committed 1 year ago

3b195fd3

Made max_pool2d_with_indices_backward_cuda contiguify `indices` (#85493)

zou3519 committed 1 year ago

f371b526

Newer Older