strisunshinewentingwang/pytorch

Pull Requests Commits

Fix header inclusions in c10 by iwyu (#100304)

cyyever committed 1 year ago

39ec5fa7

[RFC] Add tqdm to benchmarking script (#100969)

ezyang committed 1 year ago

c6587329

[FSDP][state_dict] Make sharded_state_dict work with composable fully_shard (#100856)

fegin committed 1 year ago

0fbe55ea

[PG Wrapper] Add diff capability (#100214)

rohan-varma committed 1 year ago

9ba2bfea

Revert "Fix ordered dict loading with LibTorch (#100743)"

pytorchmergebot committed 1 year ago

9ff547a5

Optimize split-split pass (#100983)

devashishshankar committed 1 year ago

cb668b12

[export] More robust view->view_copy pass (#100908)

tugsbayasgalan committed 1 year ago

f542b31c

[DataPipe] Add generated docstring to functional form DataPipe (#100503)

NivekT committed 1 year ago

8a193c6d

[opinfo] item (#100313)

khushi-411 committed 1 year ago

51fe53e6

[FSDP][state_dict] Restore the state_dict_config for NO_SHARD (#100855)

fegin committed 1 year ago

55844dfd

fix _privateuse1_tag problem (#100632)

XDaoHong committed 1 year ago

a723f1f2

[opinfo prims] equal (#100663)

khushi-411 committed 1 year ago

5a933d04

[CUDA][CUBLAS] Fix BF16 reduced precision reduction note in docs (#101044)

eqy committed 1 year ago

33f3dca6

[CUDA][CUBLAS] Add cuBLAS workspace allocation behavior to docs (#100919)

eqy committed 1 year ago

6e2efd16

Apply static policy correctly to unspec (#98983)

ezyang committed 1 year ago

1e89a56a

Fix typo in RELEASE.md and README.md (#100536)

Djacon committed 1 year ago

dfa95117

inductor(CPU): fix issue when padding/stride/dilation size is one for cpu weight packing pass (#100951)

XiaobingSuper committed 1 year ago

2b250e19

PyTorch -> C++17 (#98209) (#100557)

r-barnes committed 1 year ago

083f88e1

[MPS] Fix build regressions introduced by #92868 (#101036)

malfet committed 1 year ago

30cecc0e

[inductor] default cache_dir in torch._inductor.codecache should be lazily evaluated (#100824)

hsfzxjy committed 1 year ago

b004c0b3

CUBLAS Flag (`CUBLAS_GEMM_DFALT_TENSOR_OP` -> `CUBLAS_GEMM_DEFAULT_TENSOR_OP`) (#100976)

DyeKuu committed 1 year ago

b06c180a

[vision hash update] update the pinned vision hash (#101032)

pytorchmergebot committed 1 year ago

535368f0

[c10d] make ProcessGroupNCCL work.wait() respect timeout (#100162)

cdzhan committed 1 year ago

649e6096

[inductor] support vec type conversion between float and bool (#100950)

jgong5 committed 1 year ago

b33c9c7c

Extend assert statement to include ListVariable (#100841)

Bearnardd committed 1 year ago

44e73da4

STFT: correct stft definition and better document tensor shapes (#100427)

peterbell10 committed 1 year ago

27d5019e

Revert "[quant][pt2e] Change input act annotation to a map and allow dynamic quantization for non zeroth argument (#101005)"

pytorchmergebot committed 1 year ago

2241aaa6

[CI] Delete skips from https://github.com/pytorch/pytorch/issues/93847 (#96049)

desertfire committed 1 year ago

76cc3ab4

explicitly check or discard cudaGetLastError return value (#100488)

jeffdaily committed 1 year ago

bf214f40

[quant][pt2e] Change input act annotation to a map and allow dynamic quantization for non zeroth argument (#101005)

jerryzh168 committed 1 year ago

f08ddae8

Newer Older