strisunshinewentingwang/pytorch

Pull Requests Commits

[Easy][FSDP] Remove outdated comment (#85051)

awgu committed 2 years ago

cdd625ba

[FSDP] Fix `pin_memory()` for CPU offloading (#85048)

awgu committed 2 years ago

cc62ad79

Resubmit bfloat support for im2col,col2im (#84372)

rohan-varma committed 2 years ago

e7ad699b

Python Dispatcher integration with C++ dispatcher (#85050)

voznesenskym committed 2 years ago

8ca1839d

[functorch] fix vmapvjpvjp test for prelu (#84939)

zou3519 committed 2 years ago

3a107bc9

[functorch] fix conv_transpose with groups batching rule (#84938)

zou3519 committed 2 years ago

8badb00f

[CheckpointWrapper] Reentrant kwarg support (#84908)

rohan-varma committed 2 years ago

8cb78268

[CheckpointWrapper] Decouple CPU offload (#84907)

rohan-varma committed 2 years ago

55ca6901

[functorch] fix jacrev so all inputs get wrappers (#84914)

samdow committed 2 years ago

166ea7e6

[MPS] Fix int rounding div crash on M1 (#85016)

malfet committed 2 years ago

1a6cf6ea

[c10d] add ncclGetLastError to NCCL pg (#83724)

wanchaol committed 2 years ago

976f8bee

Don't detach when making views; force caller to detach (#84893)

ezyang committed 2 years ago

ccade941

Revert "[CUBLAS][CUDA GRAPHS] Explicitly set the workspace for cuBLAS handles (#83461)"

pytorchmergebot committed 2 years ago

2711b9fa

Fix inheritance in TestDataLoaderUtil (#85018)

ZainRizvi committed 2 years ago

a1a95d40

[CUBLAS][CUDA GRAPHS] Explicitly set the workspace for cuBLAS handles (#83461)

eqy committed 2 years ago

713d8b85

Enable manual test config label selection on GHA macos (#84895)

huydhn committed 2 years ago

7b64c885

Revert "[numpy] add `torch.concatenate`, alias of torch.cat (#82946)"

pytorchmergebot committed 2 years ago

fa7bf3e2

Shard distributed tests on non CUDA focal (#84891)

huydhn committed 2 years ago

23b7a5fc

Ensure as_strided_tensorimpl is never called with MPS (#85020)

ezyang committed 2 years ago

3e57c955

[CUDA graphs] Fixes errors in RNG seed (#84967)

Aidyn-A committed 2 years ago

5271494e

[numpy] add `torch.concatenate`, alias of torch.cat (#82946)

khushi-411 committed 2 years ago

270e5e51

Revert "Create Cache for Fusion Reuse in NVFuser in Python Frontend for Primtorch (#83267)"

pytorchmergebot committed 2 years ago

94b67f4c

[MPS] Fix mps to cpu casting from a smaller dtype to a bigger dtype (#84928)

DenisVieriu97 committed 2 years ago

4247cc98

Test tracing consecutive comms on the same input tensor (#84980)

mrshenli committed 2 years ago

1a81ab3b

Remove eager mode support form CommTensor (#84978)

mrshenli committed 2 years ago

0f300592

[ROCM] test_batchnorm_cudnn_nhwc (#84603)

micmelesse committed 2 years ago

b6d6a78c

Revert "Python Dispatcher integration with C++ dispatcher (#84826)"

pytorchmergebot committed 2 years ago

706b9903

[2/N] [Dispatchable Collectives] Extract ProcessGroup::Work into a separate class and update references (#83680)

H-Huang committed 2 years ago

74ead619

Upgrade to CUDNN version for cuda 11.7 (#84964)

atalman committed 2 years ago

54c46e4f

Skip validate_view_consistency for nvFuser tests (#84858)

IvanYashchuk committed 2 years ago

6750946b

Newer Older