strisunshinewentingwang/pytorch

Pull Requests Commits

[export] fakify module state in nonstrict (#119297)

suo committed 229 days ago

f79ae759

[AOTI][refactor] Split common aoti_runtime utils into a separate header (#119066)

desertfire committed 229 days ago

40ec155e

Migrate load_state_dict hook tests to OptimizerInfo (#119310)

janeyx99 committed 229 days ago

059994d2

Migrate test_state_dict hooks to OptimizerInfo (#119308)

janeyx99 committed 229 days ago

0320e622

[RELAND] refactor lazy init to device-agnostic (#119248)

guangyey committed 229 days ago

5c46600f

Move step global hooks test to OptimizerInfo (#119299)

janeyx99 committed 229 days ago

3625ccfb

Move step pre/post hook tests to OptimizerInfo (#119288)

janeyx99 committed 229 days ago

7b3762e6

Add symbol guard counts instrumentation (#119290)

ezyang committed 229 days ago

99ddfaf5

Add basic reference documentation for symbolic_shapes.py (#118997)

peterbell10 committed 229 days ago

7c95cc5e

Increase accumulate_grad_ gradient's expected refcount to account for pybind (#119068)

xmfan committed 229 days ago

1435cfec

Never reuse accumulated gradients' buffers (#119334)

xmfan committed 229 days ago

326dcf9d

Fix gradient refcounts in pybind and compiled autograd (#118817)

xmfan committed 229 days ago

8e14e1d5

Revert "Fix deadlock in ExecutionTraceObserver (#119242)"

pytorchmergebot committed 229 days ago

d85631b7

add Half support for flash attention (#119247)

CaoE committed 229 days ago

dfdbd733

[Dynamo][15/N] Merge allow_in_graph/inline/skip trace rules check into trace_rule.lookup (#118971)

yanboliang committed 229 days ago

0f478d9d

Add --local-ranks-filter to torchrun: allow logs filtering by rank (#118562)

xmfan committed 229 days ago

284b0b5f

Enable optional tensorList fallback to cpu. (#119273)

Shan19900305 committed 229 days ago

6c3600d0

[vision hash update] update the pinned vision hash (#119337)

pytorchupdatebot committed 229 days ago

53ee47ca

[dynamo] delete dynamo cache entry when guard function is invalidated [attempt 2] (#119107)

williamwen42 committed 229 days ago

ee1c2449

[ONNX][dynamo_export] Turn off opmath type promotion for div (#119112)

BowenBao committed 229 days ago

fcc36de9

Add torch.dtype instances to the public API (#119307)

TamirCohen committed 229 days ago

45a79323

[EZ][BE] [CMake] Remove checks for GCC-7 (#119306)

malfet committed 229 days ago

8c2fde1f

[PyTorch] Free up 8 bytes per intrusive_ptr_target (#117986)

swolchok committed 229 days ago

e9907a34

Fix typo on torch.frombuffer() documentation (#119214)

mdevino committed 229 days ago

5f2ad407

Test seo torch cuda (#119324)

svekars committed 229 days ago

5ae6f6cf

LazyGraphModule: improve the fix for the FakeTensorMode mismatch issue (#119311)

shunting314 committed 229 days ago

728228a7

[AOTI] Rename config.aot_inductor.abi_compatible (#119065)

desertfire committed 229 days ago

e868a7fe

Fix handling random() calls encountered inside inlined code. (#119218)

laithsakka committed 229 days ago

c814d8e5

[dynamo] Functools partial reconstruct (#118583)

jansel committed 229 days ago

5e78c4b0

[dynamo] Fix missing guards in FunctoolsPartialVariable (#118616)

jansel committed 229 days ago

62cc1053

Newer Older