strisunshinewentingwang/pytorch

Pull Requests Commits

Fix `ShardedTensorMetadata.tensor_properties` for Python 3.11 (#91795)

ConnorBaker committed 1 year ago

20665235

[threaded_pg] fix the comments of MultiThreadTestCase (#92373)

wanchaol committed 1 year ago

06d54b40

[dtensor] delete lagging op db and update op db tests (#92290)

wanchaol committed 1 year ago

997de441

Improve `bsr @ strided` performance in `baddmm` for `bfloat16/half` with Triton kernels. (#88078)

nikitaved committed 1 year ago

8383b5c4

some fixes to get symbolic shapes working through inductor (#92320)

Chillee committed 1 year ago

4f4b62e4

Fix key error formatting and move exc code to exc.py (#92593)

mlazos committed 1 year ago

cac217c8

Make run_dynamic_ci_skips_only.sh more generic (#92581)

ezyang committed 1 year ago

ba682057

[CI] move parallelnative to periodic (experimental) (#92567)

janeyx99 committed 1 year ago

2a7a859d

Remove temporary export skip hack (#92160)

voznesenskym committed 1 year ago

28cb3141

fix promote_constants with ExpandView (#92403)

Natalia Gimelshein committed 1 year ago

ef258642

When nopython=True, Dynamo can't allow graph breaks. (#90970)

wschin committed 1 year ago

bdbd3ed3

Guard on at::Tensor device index (#91779)

voznesenskym committed 1 year ago

eb39d990

[CI] valgrind 3.16.1->3.20.0 (#92552)

malfet committed 1 year ago

388d79cc

Make aot_autograd explicitly error when double backward (#92348)

soulitzer committed 1 year ago

bb779078

[PTD][Oncall] Sync Reorder structure for compatibility with linux-6.0 and gloo submodule for PT (#92568)

fduwjj committed 1 year ago

62eeb7d6

[mergebot] Flatten workflows into jobs, fix bugs (#92097)

clee2000 committed 1 year ago

34353a40

Remove lowering and decompositions of zero_, zero, zeros_like... in favour of their references (#92071)

lezcano committed 1 year ago

8b861544

Fix dynamo.export(aten=True) for condition op (#92361)

SherlockNoMad committed 1 year ago

b5c3b4a3

[optim][asgd] group tensors in foreach to maximize perf (#92364)

janeyx99 committed 1 year ago

c5cb46ec

Fix checking of current mode in PyOperator dispatch (#92357)

angelayi committed 1 year ago

5fdddbbf

Reland 2nd attempt "Add heirachical module names to torchFX graph.node" (#91721)

alexmsettle committed 1 year ago

f8a07ca4

fix incorrect _embedding_bag meta (#92549)

bdhirsh committed 1 year ago

76cb2d0e

Change references to pytorch/functorch to the torch.func APIs (#92543)

zou3519 committed 1 year ago

5aa3740d

[optim][radam] group tensors in foreach to maximize perf (#92365)

janeyx99 committed 1 year ago

fbafcecf

[optim][rmsprop] group tensors in foreach to maximize perf (#92369)

janeyx99 committed 1 year ago

de459bdf

[optim][adam] group tensors in foreach to maximize perf (#92349)

janeyx99 committed 1 year ago

07800c52

[optim][adamax] group tensors in foreach to maximize perf (#92363)

janeyx99 committed 1 year ago

e2433e42

[FSDP][optim_state_dict][11/N] Let FSDP support NamedOptimizer/KeyedOptimizer when use_orig_params is False (#92184)

fegin committed 1 year ago

92d412d6

Revert "Clean up C++14 code (#92216)"

pytorchmergebot committed 1 year ago

befe3b68

Reduce some ambiguity in Tensor (#92266)

Yinghai Lu committed 1 year ago

4450424b

Newer Older