strisunshinewentingwang/pytorch

Pull Requests Commits

zou3519 committed 2 years ago

92fc990e

voznesenskym committed 2 years ago

663f8ca0

Feedback pass, cleanup, improvements

voznesenskym committed 2 years ago

50181ee9

Feedback pass 3

voznesenskym committed 2 years ago

366591c3

Feedback pass 2

voznesenskym committed 2 years ago

2b7b8760

voznesenskym committed 2 years ago

2059523a

Add back accidently rm important line

voznesenskym committed 2 years ago

ee61ade8

Make it less bad, safeguards, questionable choices

voznesenskym committed 2 years ago

52ffdc6f

Merge branch 'master' of github.com:pytorch/pytorch into voz/ctfl_proto

voznesenskym committed 2 years ago

1bde5ee4

Enable pg_nccl to perform vector AllGather for uneven output splits (#83713)

aashaka committed 2 years ago

30742193

[Profiler] Start moving python bindings out of autograd (#82584)

Taylor Robie committed 2 years ago

1fa9a377

Remove duplicate_dequantize_node and remove_extra_dequantize (#83611)

danrecoskie committed 2 years ago

7453019e

[Vulkan] Fix issues in GRU and LSTM (#83722)

manuelcandales committed 2 years ago

da520a43

Avoid using fx.Interpreter in nvfuser executor function (#83607)

IvanYashchuk committed 2 years ago

108a1fb1

[vulkan] Throw std::runtime_error instead of using TORCH_CHECK when creating Vulkan context/runtime fails (#83627)

SS-JIA committed 2 years ago

e0d26ee0

Nvfuser python api patch take 2 (#83684)

jjsjann123 committed 2 years ago

1407e672

Refactor CppSignatureGroup to collect signatures as list. (#83667)

ezyang committed 2 years ago

0ec7fc13

Switch fx.replace_pattern to use new SubgraphMatcher (#83717)

SherlockNoMad committed 2 years ago

03e322c8

Avoid unnecessary copy of pointeeSet in MemoryDAG::setWildcards (#83681)

ezgicicek committed 2 years ago

73652dd1

[functorch] re-classify linalg.eigh in vmap testing (#83614)

zou3519 committed 2 years ago

93eedc51

Move `torch.linalg` opinfos to opinfo.definitions (2/2) (#83554)

peterbell10 committed 2 years ago

8788e92f

Move `torch.linalg` opinfos to opinfo.definitions (1/2) (#83547)

peterbell10 committed 2 years ago

8dbb0990

Move RefInfo classes into opinfo.refs (#83563)

peterbell10 committed 2 years ago

4aeb98de

Move gradcheck_wrapper and clone_sample funcs into opinfo.core (#83560)

peterbell10 committed 2 years ago

f4caeb25

Enable formatting in all of testing/_internal/opinfo (#83559)

peterbell10 committed 2 years ago

ae68e455

[composite compliance] batch_norm (#79990)

kshitij12345 committed 2 years ago

b4bc0d24

Ensure cuda_primary_ctx test is run on multigpu CI (#83252)

lw committed 2 years ago

a6f777c8

[vision hash update] update the pinned vision hash (#83729)

pytorchmergebot committed 2 years ago

ca9919e3

Revert "Manually shard slow-gradcheck CI job to prevent timeout #83354" (#83704)

clee2000 committed 2 years ago

b8d647e1

Revert "feedback pass"

voznesenskym committed 2 years ago

9de29729

Older