strisunshinewentingwang/pytorch

Pull Requests Commits

[c10][rocm] fix __assert_fail() declaration mismatch error (#73040)

shintaro-iwasaki committed 2 years ago

7dc2cfa2

Back out "[BE][DDP] enable rebuilt bucket when find_unused_parameters=True" (#73524)

zhaojuanmao committed 2 years ago

6b883d99

Back out "[DDP][BE] Remove bucket replicas" (#73523)

zhaojuanmao committed 2 years ago

a9c1c205

Back out "[DDP][BE] Fix clang-tidy" (#73522)

zhaojuanmao committed 2 years ago

fb6977cb

Back out "[Easy][c10d] Minor fixes" (#73521)

zhaojuanmao committed 2 years ago

2d45a3d7

[dnnlowp] fix qparam handling during bias quantization (#73478)

jspark1105 committed 2 years ago

52b707b5

[torch] set workspace size for cublas lt interface 1M (#73439)

jspark1105 committed 2 years ago

3ec1dd99

[Static Runtime] [Code Cleanup] Use `SROperator` for operators' function type (#73450)

d1jang committed 2 years ago

c62de0ac

Move util functions to a more common place (#73519)

842974287 committed 2 years ago

5e865056

[quant] Enable reference path by default for CopyNodeQuantizeHandler (#73233)

jerryzh168 committed 2 years ago

ad1078a2

Disable cublasLt when batch is too large. (#73533)

jasonjk-park committed 2 years ago

31d742b6

[ONNX] Export bias requantize steps to ONNX

BowenBao committed 2 years ago

8280919f

ufunc codegen (#65851)

ezyang committed 2 years ago

ce7910ba

Disable TF32 in some linalg functions (#73460)

xwang233 committed 2 years ago

89b4cfb4

codegen/templates: disable clang-format check (#73057)

guoyejun committed 2 years ago

78914b30

pytorch: fix typo in quantization docs (#73511)

vkuzo committed 2 years ago

01bd6f43

[PyTorch Distributed] Fix NCCL version string (#73333)

kwen2501 committed 2 years ago

1e7a4d6b

[Core] Update Exception.h (#64553)

makslevental committed 2 years ago

20a037d8

Enable test_coalesce_cuda_bfloat16 (#73158)

pearu committed 2 years ago

a5dcc0c3

Support dataclasses in TorchScript (#73066)

norabelrose committed 2 years ago

d00de0d4

[ONNX] Capture annotated attributes for local function

BowenBao committed 2 years ago

b3cfc74f

[Static Runtime] Add test that runs NNC fused kernels in parallel (#73256)

navahgar committed 2 years ago

cfd92f2d

Add api for recursively analyzing function calls (#73329)

Elias Ellison committed 2 years ago

ab6395fc

Allow custom registration of shape functions (#73270)

Elias Ellison committed 2 years ago

d3d74e90

Enable win-arm64

gaborkertesz-linaro committed 2 years ago

c4ff49f4

[BE][GHA] Use `timeout_after` for win templates

malfet committed 2 years ago

a1d5b5d2

fused int8 static (#73452)

xiaosun86 committed 2 years ago

14365079

Add pickling support for WorkerInfo (#73371)

H-Huang committed 2 years ago

6c8e516a

Remove legacy tensor constructors for complex dtypes

peterbell10 committed 2 years ago

f437ca6e

Refactor set_default_tensor_type to avoid legacy tensor types

peterbell10 committed 2 years ago

d51103e7

Newer Older