strisunshinewentingwang/pytorch

Pull Requests Commits

Autorun binary builds when a commit pin is updated (#96526)

ZainRizvi committed 1 year ago

19833486

Add merge_rule for "functorch" module (#96657)

zou3519 committed 1 year ago

7eef4697

[PT-D] Update CODEOWNERS, merge_rules, and Persons-of-Interest for to… (#96321)

kiukchung committed 1 year ago

55a1bd3f

Dockerize torch deploy setup (#96593)

huydhn committed 1 year ago

bb8dc7f7

sparse_mask: remove syncs by removing calls to coalesce (#94406)

nikitaved committed 1 year ago

0b5040b3

Fix vmap registration for t, t_ (#96539)

zou3519 committed 1 year ago

13011afb

Support zeros_like() for NT (#96527)

jbschlosser committed 1 year ago

024ea1a2

Corrected `HingeEmbeddingLoss` documentation (#95140)

ZiadAmerr committed 1 year ago

3cdf18cb

DDP native mixed precision (#92882)

rohan-varma committed 1 year ago

32f11f58

Update CI skips (#96554)

ezyang committed 1 year ago

c7f39c08

inductor(cpu): make variable number used of masked vectorization path align with scalar path (#96510)

XiaobingSuper committed 1 year ago

279ada51

Enablee test_inverse_errors_large (#94727)

mantaionut committed 1 year ago

2cbce06f

[Dynamo] User defined functions support torch & builtin functions as default arguments (#96563)

yanboliang committed 1 year ago

760ad905

inductor(cpu): remove __restrict__ keyword to avoid generating wrong result when two pointer point same memory (#96492)

XiaobingSuper committed 1 year ago

6eca391e

Revert "[primTorch] add meta implementation for `upsample_nearest2d_backward` (#96612)"

pytorchmergebot committed 1 year ago

be220690

Deprecate non-bool masks in masked_fill (#96594)

kvathupo committed 1 year ago

2b9d9bcb

[primTorch] add meta implementation for `upsample_nearest2d_backward` (#96612)

nkaretnikov committed 1 year ago

fe180596

Generate type match guard for torch.Size input (#96421)

ezyang committed 1 year ago

99efe3ef

[BE] Dedup hardcoded triton versions (#96580)

malfet committed 1 year ago

1ab88379

Revert "[BE] Dedup hardcoded triton versions (#96580)"

pytorchmergebot committed 1 year ago

30b968f6

[BE] Dedup hardcoded triton versions (#96580)

malfet committed 1 year ago

c131e51e

[memory profiling] C++ tracing support (#95357)

zdevito committed 1 year ago

4b372e39

[memory profiling] Move Context object to c10 (#96280)

zdevito committed 1 year ago

48490cec

[memory snapshots] record scripted stack traces (#95356)

zdevito committed 1 year ago

266089a3

Fix unpicklable object in AveragedModel (#95979)

DavidFM43 committed 1 year ago

e8b0f504

Properly capturing argument names for decorated/wrapped functions (#96557)

gmagogsfm committed 1 year ago

82d3d053

Add location information for assertions in `torch.jit.annotations.try_ann_to_type` (#96423)

FabianSchuetze committed 1 year ago

a7a09adb

Symintify `_gather_sparse_backward` (#96591)

nkaretnikov committed 1 year ago

12735952

Enable `min.unary_out` (#96441)

nkaretnikov committed 1 year ago

cb7c796b

[pt2][inductor] Ignore trace.upload_tar when pickling config (#96519)

bertmaher committed 1 year ago

31a67304

Newer Older