strisunshinewentingwang/pytorch

Pull Requests Commits

[MPS] Fix torch.full for uint8 (#83697)

malfet committed 2 years ago

ff533b1e

Fix and improve the efficiency of the backward of xlog* functions. (#82713)

lezcano committed 2 years ago

88d3acd6

[primTorch] Refs for pdist, triu and related ops (#82819)

Fabio Rocha committed 2 years ago

9e156082

[DataLoader] Move loop content into a function to ensure we don't preserve anything (#83595)

albanD committed 2 years ago

38348362

Add remaining device types in the pybinded DeviceType enum (#83676)

H-Huang committed 2 years ago

23d22724

Revert "Remove conj kernels for real dtypes (#80374)"

pytorchmergebot committed 2 years ago

46ba9f2e

[dist.cp] Introduce LoadPlanner and SavePlanner extensibility API. (#83419)

Rodrigo Kumpera committed 2 years ago

d11d3dd0

[functorch] reclassify svd as an allowed failure; add test (#83612)

zou3519 committed 2 years ago

4a033be4

[functorch] add some vmap+jvp inplace+view tests (#83178)

zou3519 committed 2 years ago

601aca2a

[functorch] relax as_strided batching rule (#83597)

zou3519 committed 2 years ago

d84dc589

[functorch] annotate test_jvpvjp (#83530)

zou3519 committed 2 years ago

69728d7d

[ONNX] Update the default opset to version 14 (#83284)

justinchuby committed 2 years ago

e4f74f08

Added communication hook for sharded cases (#83254)

Olga Andreeva committed 2 years ago

f204afc2

[quant][ao_migration] `torch.nn.quantized.functional` → `torch.ao.nn.quantized.functional` (#78712)

zaf committed 2 years ago

78c8a0d7

[FSDP] Implement sharded_optim_state_dict and flatten_sharded_optim_state_dict. (#77628)

fegin committed 2 years ago

3e1fc85b

Handle python frame is empty in GetPythonFrames (#83643)

JackCaoG committed 2 years ago

cd0ab154

Release the GIL when munmap'ing tensors - fixes #77139 (#83623)

cchan committed 2 years ago

abcf0119

Revert "fixing define_constant pybind signature to match std::complex scalar (#83645)"

pytorchmergebot committed 2 years ago

f84e087d

fix quantization/core/test_docs for Buck2 (#83341)

mikey dagitses committed 2 years ago

3f612b58

Make the derivative of masked_fill more efficient (#83515)

lezcano committed 2 years ago

aad89bb7

[vision hash update] update the pinned vision hash (#83582)

pytorchmergebot committed 2 years ago

4b3f1bdb

[xla hash update] update the pinned xla hash (#83581)

pytorchmergebot committed 2 years ago

eb600414

[MPS] Register index.Tensor_out (#82507)

kulinseth committed 2 years ago

ce7177f8

Update ideep for NNC post-op (#82705)

yanbing-j committed 2 years ago

6dc8673b

fixing define_constant pybind signature to match std::complex scalar (#83645)

jjsjann123 committed 2 years ago

278c7264

[torchgen] Relax the restriction on number of custom namespaces (#83580)

larryliu0820 committed 2 years ago

badbdb03

Revert "[primTorch] Add ref for `new_empty_strided` (#82466)"

pytorchmergebot committed 2 years ago

7263450c

Enable pg_nccl.reduce_scatter to perform vector ReduceScatter for uneven input splits (#82924)

aashaka committed 2 years ago

d6a30e21

Delete unnecessary sum.SymInt overload (#83591)

ezyang committed 2 years ago

52be9082

SymInt'ify schemas for prims (#83528)

ezyang committed 2 years ago

6679d238

Newer Older