strisunshinewentingwang/pytorch

Pull Requests Commits

[Reland][functorch] test for compiling functorch transforms (#100718)

zou3519 committed 1 year ago

d08066a4

Add mechanism to turn any RAII guard into a Python Context Manager (#102037)

zou3519 committed 1 year ago

08fb648f

determined collective device by _get_pg_default_device rather than explicit cuda (#101533)

medivh-xp committed 1 year ago

8b7bd819

[inductor] Add more dynamic shapes support for CudaWrapperCodeGen (#102019)

desertfire committed 1 year ago

fd1d4421

[c10d] Record time spent for init_process_group, new_group, _store_based_barrier (#101912)

wz337 committed 1 year ago

ee95e37a

add Half support for sinh, cosh, ploygamma, entr and i0e on CPU (#99002)

jiayisunx committed 1 year ago

d6afa7d0

Bump mpmath from 1.2.1 to 1.3.0 in /.github/requirements (#102058)

dependabot[bot] committed 1 year ago

8aea9dad

change error_message for XPU Autocast data type check (#102073)

xiaolil1 committed 1 year ago

faa7eb81

No need to run C++ tests under rerun disabled tests mode (#102132)

huydhn committed 1 year ago

d0680277

Enable mypy allow redefinition (#102046)

ringohoffman committed 1 year ago

29da75cc

[Typing] Export `torch.backends` as subpackage (#102099)

malfet committed 1 year ago

bf059e39

Lower xdist processes from auto to NUM_PROCS (#102124)

huydhn committed 1 year ago

d26c8f26

Tighten FakeTensor reentrancy asserts, add debugging (#102091)

ezyang committed 1 year ago

3318a832

group constraints by arg (#102096)

avikchaudhuri committed 1 year ago

38f8f756

[vision hash update] update the pinned vision hash (#102136)

pytorchupdatebot committed 1 year ago

907cc6c1

Improve bf16 neg by bypassing the convertion between BF16 and FP32 (#99711)

EikanWang committed 1 year ago

2e18dd2b

test_memory_format fix for test_modules.py (#102006)

Fuzzkatt committed 1 year ago

45843c7f

move tf32_on_and_off fix for test_convolution.py (#102007)

Fuzzkatt committed 1 year ago

47e9dba7

disable tf32 for rnn tests and norm tests (#102005)

Fuzzkatt committed 1 year ago

d805a53f

Remove config check in specialize (#102098)

voznesenskym committed 1 year ago

ea5eaa86

Disable (Broken) CUDAStreamVariable in dynamo (#100766)

voznesenskym committed 1 year ago

d55aad1f

integrate the new event with pytorch (#101025)

fenypatel99 committed 1 year ago

cc233f4e

[pt2] add `SymInt` support for `linalg.matrix_power` (#101940)

nkaretnikov committed 1 year ago

e79d9b99

[pt2] add `SymInt` support for `eye` (#101955)

nkaretnikov committed 1 year ago

69f7b409

[pt2] add meta for `linalg_lu_solve` (#101836)

nkaretnikov committed 1 year ago

42b974e8

[MPS] Calculate nonzero count first before running nonzero op (#102052)

DenisVieriu97 committed 1 year ago

6c681166

Make _StorageBase.byteswap faster ( > 10000x) (#101925)

kiszk committed 1 year ago

be5e77ca

[quant][pt2e] prepare_pt2e use quantization spec directly (#102054)

jerryzh168 committed 1 year ago

94ed26d1

[PyTorch] Delete c10::guts::if_constexpr (#101991)

swolchok committed 1 year ago

99f68d56

Support FakeTensor with FlatParameter (#101987)

ezyang committed 1 year ago

f6573255

Older