strisunshinewentingwang/pytorch

Pull Requests Commits

Revert "Update jvp to support symbolic execution. (#120338)"

pytorchmergebot committed 183 days ago

88999674

[codemod] Remove unused variables in caffe2/caffe2/experiments/operators/fully_connected_op_prune.h (#122165)

r-barnes committed 183 days ago

e0d57001

release gil in prepareProfiler (#121949)

ngimel committed 183 days ago

6bd2d12b

[PT2] - Fix cat backwards wrapping on symints (#121527)

flaviotruzzi committed 183 days ago

7fb2d692

Back out "[fx] Preserve Fx graph node order in partitioner across runs (#115621)" (#122113)

huydhn committed 183 days ago

8de4d864

[torch export][serialize] create a more compact stacktrace format for serialization (#121675)

strisunshinewentingwang committed 183 days ago

eae89138

[Functorch] Bump tolerances for `test_per_sample_grads_embeddingnet_mechanism_functional_call_cuda` (#122014)

eqy committed 183 days ago

271b12c7

Add scuba logging for TorchScript usage (#121936)

gmagogsfm committed 183 days ago

ba9a1d96

[TD] Add LLM retrieval + heuristic (#121836)

clee2000 committed 183 days ago

4819da60

[pt2] add symbolic shape support for decompose mm and expose max_block to user config (#121440)

jackiexu1992 committed 183 days ago

cec0fd6f

Revert "Add Flash Attention support on ROCM (#121561)"

pytorchmergebot committed 183 days ago

764eae9c

[dynamo] Forward OptimizedModule.__setattr__ to the wrapped module (#122098)

peterbell10 committed 184 days ago

88ebdbc9

Flatten/Unflatten micro optimization in proxy_tensor.py (#121993)

mlazos committed 184 days ago

2164b7f7

Fixes nan with large bf16 values (#122135)

drisspg committed 184 days ago

42624bce

Fix typing for autograd.Function with ctx-less forward (#122167)

zou3519 committed 184 days ago

e26280ad

Revert "Refactor gpu trace to be device-agnostic (#121794)"

pytorchmergebot committed 184 days ago

f9ed1c43

[dynamo] Remove copy_graphstate/restore_graphstate (#122067)

jansel committed 184 days ago

c05bf003

Revert "Skip nonzero unbacked SymInt memo in inference mode (#122147)"

pytorchmergebot committed 184 days ago

7673cb53

[Clang-tidy header][28/N] Fix clang-tidy warnings in aten/src/ATen/core/*.h (#122175)

cyyever committed 184 days ago

6c01c253

[ATen-Vulkan][EZ] Small fixes: fix gpu size calculation and Half scalartype ctype mapping (#122096)

SS-JIA committed 184 days ago

6c503088

Update jvp to support symbolic execution. (#120338)

guilhermeleobas committed 184 days ago

39877abe

Teach dynamo about torch.func.jvp (#119926)

guilhermeleobas committed 184 days ago

edd04b7c

[lint] bump lint dependency PyYAML to 6.0.1 to support Python 3.12 (#122022)

XuehaiPan committed 184 days ago

6b5259e5

Add CPU implementation for `torch._int_mm` (s8*s8->s32) (#121792)

Xia-Weiwen committed 184 days ago

81683380

Fix auto_functionalize (#121990)

tugsbayasgalan committed 184 days ago

0d845f7b

Avoid COW materialize in conv, log sigmoid, repeat, group_norm, batch_norm (#121537)

kurtamohler committed 184 days ago

a2a88f39

Refactor gpu trace to be device-agnostic (#121794)

guangyey committed 184 days ago

0ff1109e

Improve compiler detection on MacOS (#121406)

xuhancn committed 184 days ago

09ce7680

add sdpa choice for DeviceType::PrivateUse1 (#121409)

1274085042 committed 184 days ago

8499767e

[dynamo] Make tx.next_instruction lazy (#122066)

jansel committed 184 days ago

5bc7f7f9

Newer Older