strisunshinewentingwang/pytorch

Pull Requests Commits

[easy] Remove duplicate exprs in produce_guards (#111270)

masnesral committed 339 days ago

ff432c04

fix: reset prefetch flag upon reshard (#111354)

0x804d8000 committed 339 days ago

b691d090

[ONNX] Fix aten::new_zeros due to TorchScript behavior change on Pytorch 2.1 Fix #110935 (#110956)

Thiago Crepaldi committed 339 days ago

9ab6ac5b

[dynamo] make disable_cahce_limit also disable accumulated cache limit (#111334)

ydwu4 committed 339 days ago

9f562a3d

Revert "[inductor] Adding a way to force fusion of int_mm with mul (#111125)"

pytorchmergebot committed 339 days ago

89f11c69

[tp] fix SP style regression (#111353)

wanchaol committed 339 days ago

59281d56

Revert "[C10D] Introduce C++ side Collective Callbacks. (#110307)"

pytorchmergebot committed 339 days ago

493618d7

Fixes a typo in docstring: should be "elastic" (#111352)

sighingnow committed 339 days ago

6462d71c

fix wrong meta for index_select.out (#111364)

bdhirsh committed 339 days ago

0d368f58

FunctionalTensor: avoid spurious not_implemented logging during proxy tracing (#111040)

bdhirsh committed 339 days ago

4cf23c6a

fix bugs about traceback.walk_stack in python3.8.x (#110922)

FFFrog committed 339 days ago

50b80185

Error if you try to run Dynamo compiled function under torch.jit.trace (#111321)

ezyang committed 339 days ago

126d422c

[xla hash update] update the pinned xla hash (#111360)

pytorchupdatebot committed 339 days ago

78909a6f

Revert "[vision hash update] update the pinned vision hash (#111316)"

pytorchmergebot committed 339 days ago

9af82fa2

Revert "[sparse] semi-structured sparse + torch.compile support (#111049)"

pytorchmergebot committed 339 days ago

b4745d47

[TP] Fix TP doc format to show examples correctly (#111346)

fduwjj committed 339 days ago

bfcd8695

update int4 tinygemm kernels (#111327)

Chillee committed 340 days ago

e0e15a4a

[dtensor][11/n] adds some __str__ for ease of read (#111278)

wanchaol committed 340 days ago

882bc170

[dtensor][10/n] switch pointwise op to use op strategy (#111234)

wanchaol committed 340 days ago

6b5d736b

[dtensor][9/n] matrix ops to generate strategy (#110717)

wanchaol committed 340 days ago

f34f3b54

[dtensor][8/N] Introduce cost model for sharding (#109145)

wanchaol committed 340 days ago

b4ab8ac5

[TP] Enable embedding sharding in TP API (#111177)

fduwjj committed 340 days ago

25a2845d

Fix get_estimated_runtime for symbolic shapes (#111314)

Chillee committed 340 days ago

e942fddb

direct runtime assertions (#111262)

avikchaudhuri committed 340 days ago

e6d9350d

[state_dict][4/N] Support strict flag for model.load_state_dict (#111109)

fegin committed 340 days ago

7df287dc

[vision hash update] update the pinned vision hash (#111316)

pytorchupdatebot committed 340 days ago

da364449

Update mypy to 1.6.0 (#111305)

malfet committed 340 days ago

4a388e70

trace frames with np.ndarray (#110512)

ev-br committed 340 days ago

48989bc8

[Dynamo] Fix autograd.Function tracing errors loudly involving saved tensors (#111277)

yanboliang committed 340 days ago

da662248

[TP] Add deprecation warnings in the documentations for Pairwise parallel, sequence parallel and other prepare input/output functions (#111176)

fduwjj committed 340 days ago

ff3d773d

Newer Older