strisunshinewentingwang/pytorch

Pull Requests Commits

[Dynamo] Fix bug when dynamo generate guards for enum type (#98652)

yanboliang committed 1 year ago

a5f34686

reenable lowmem dropout (#98631)

ngimel committed 1 year ago

0dbdc8a3

distinguish mutability of untyped Storage::data (#97690)

mikey dagitses committed 1 year ago

c68a94c5

Add NLLLoss to DTensor prop rule (#98512)

mrshenli committed 1 year ago

d255c8e1

Set up automated hash pinning for triton (#97568)

clee2000 committed 1 year ago

a6155f34

Modify 'fake_tensor_unsupported' function (#98585)

jiayisunx committed 1 year ago

f959a0d5

[inductor] Use 64-bit indexing for large tensors in triton code (#97447)

peterbell10 committed 1 year ago

b7ff7172

[inductor] Fix benchmark_compiled_module codegen with CppWrapperCodeGen (#98608)

peterbell10 committed 1 year ago

48397cdd

Fix pytest config (#98607)

peterbell10 committed 1 year ago

917e9f11

Update the use of nvidia-smi for GPU healthcheck (#98036)

huydhn committed 1 year ago

4563adac

Back out "[kineto] add SOFT_ASSERT when logging metdata" (#98630)

christ1ne committed 1 year ago

112dfa14

[Dynamo] Include UserDict in clone_inputs (#97725)

YJ Shi committed 1 year ago

5ceae85f

Add comments about maybe_guard usage in Inductor (#98563)

ezyang committed 1 year ago

cf10fd82

Back out "`GradScaler` recomputes `optimizer_state["found_inf_per_device"]` before `optimizer.step` (#97415)" (#98613)

apeeyush committed 1 year ago

ebd4c165

[fx] Subgraph rewriter matching on attributes (#98604)

angelayi committed 1 year ago

2d9f482d

Enable flatbuffer tests properly. (#98363)

qihqi committed 1 year ago

4adae2d1

Retry ONNX tests (the quick way) (#98627)

huydhn committed 1 year ago

0a0f107b

[ONNX] Enable xdoctests in CI (#98546)

BowenBao committed 1 year ago

4f9dbc17

Fix wrong SPMD test target in test_log_softmax (#98610)

mrshenli committed 1 year ago

b2b783ea

Fix MPI rank and world size pg initialization (#98545)

H-Huang committed 1 year ago

61c74ab0

Move functional collectives implementation to python. (#98595)

Rodrigo Kumpera committed 1 year ago

24d90015

grab bag of changes (#98572)

Chillee committed 1 year ago

c75dd7c4

Remove MERGE_IN_PROGRESS when exiting merge (#98611)

malfet committed 1 year ago

9667f261

Revert "[experiment] More procs in CI (#98098)"

pytorchmergebot committed 1 year ago

55724a5e

[CI] Mark vision_maskrcnn as NONDETERMINISTIC (#98570)

desertfire committed 1 year ago

5210d7c4

[quant][pt2e] Add support for a few ops in QNNPackQuantizer to enable quantizing internal model (#98560)

jerryzh168 committed 1 year ago

c5269ad6

Work around CI worker gpu issue for inductor_distributed (#98601)

wconstab committed 1 year ago

89e57744

[pt2] add meta functions for `cummax` and `cummin` (#98552)

nkaretnikov committed 1 year ago

1c226f5a

[Quant] Add get_symmetric_qnnpack_qat_qconfig_mapping (#98569)

mcr229 committed 1 year ago

483fd335

Clean up compile reason logic, report only graph break compiles (#98574)

voznesenskym committed 1 year ago

e016dec6

Newer Older