strisunshinewentingwang/pytorch

Pull Requests Commits

fix functionalization <> fake tensor mode (#83701)

bdhirsh committed 2 years ago

7fe19c03

reinplacing pass fixes for torchbench + huggingface (#83626)

bdhirsh committed 2 years ago

e9e73638

functionalization: handle models that resize their program inputs (#83542)

bdhirsh committed 2 years ago

cce32c6f

Always allow tensor metadata changes (#83590)

bdhirsh committed 2 years ago

0c24af49

[vulkan][ez] lock cache mutex when purging for ShaderCache (#83738)

SS-JIA committed 2 years ago

a7d8863c

Pin sphinxcontrib.katex to 0.8.6 (#83774)

huydhn committed 2 years ago

155343ef

Make some optimizations to minifier (#83641)

Chillee committed 2 years ago

2efbdbfc

[quant][fx][refactor] Rename qconfig_utils.py to qconfig_mapping_utils.py in torch/ao/quantization/fx (#83369)

jerryzh168 committed 2 years ago

13f42069

[primTorch] Add ref for `new_empty_strided` (#82466)

Nikita Karetnikov committed 2 years ago

1f38225b

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

Remove assertEqualIgnoreTypes from test_unary_ufuncs (#83711)

schlagercollin committed 2 years ago

5bc85fcc

Newer Older