strisunshinewentingwang/pytorch

Pull Requests Commits

[quant][embedding qat] Enable quint4 in EmbeddingBag QAT workflow (#66348)

b-koopman committed 2 years ago

8e27aa52

Fix `nn.functional.max_poolNd` dispatch (for arg: `return_indices`) (#62544)

krshrimali committed 2 years ago

08a423ae

Deduplicate codegenOutputQuery to query maximum CUDA compute capabilities (#55901)

Flamefire committed 2 years ago

18d8dfba

Adding StreamWrapper to ensure file object will be closed (#66715)

NivekT committed 2 years ago

b2521822

Add relative and absolute tolerances for matrix_rank, pinv (#63102)

IvanYashchuk committed 2 years ago

c528a948

[PyTorch] Allow override for macro `HAS_DEMANGLE` (#66540)

larryliu0820 committed 2 years ago

4adca5d6

Revert D31577488: Migrate THCState to ATen

ngimel committed 2 years ago

c96486ea

Revert D31547709: Remove THCGeneral.cpp

ngimel committed 2 years ago

dc4d8bda

Implement Python Array API `asarray` function. (#60627)

ysiraichi committed 2 years ago

8ea18269

Make aotCompile support multiple input sizes (#66727)

priyaramani committed 2 years ago

c40acbaf

Refactor: move method to func compilation work to compileMethod, add option to specify method name (#66726)

priyaramani committed 2 years ago

cc06abd9

Remove THCGeneral.cpp (#66391)

peterbell10 committed 2 years ago

8396e186

add frozen_numpy as a builtin library to torch::deploy (#66297)

shunting314 committed 2 years ago

c13c8ea4

[TS] Return early if device doesn't match (#66694)

842974287 committed 2 years ago

29ee50a2

Suppress deprecated copy in vec256_qint.h (#66646)

r-barnes committed 2 years ago

a7deee79

Migrate THCState to ATen (#66480)

peterbell10 committed 2 years ago

9e3c31b4

Revert D30652629: use irange for loops

Xue Li committed 2 years ago

8a618ec7

sort_out_cuda: Use custom kernels to fill index tensors (#66668)

peterbell10 committed 2 years ago

12652ae4

Clean up test running scripts (#65508)

driazati committed 2 years ago

2ff067b7

[Vulkan] cat operator for channel dimension (#66669)

beback4u committed 2 years ago

1d0c91e6

Promote integral inputs to floating for `torch.logsumexp` (#63393)

Kiyosora committed 2 years ago

135b0d20

[Static Runtime] Determine function for `ProcessedNode::run()` statically (#66692)

d1jang committed 2 years ago

071eedf7

Revert D31656999: Add meta support to tensor range factories

malfet committed 2 years ago

c2c42560

Add `warn_only` kwarg to `use_deterministic_algorithms` (#66233)

kurtamohler committed 2 years ago

79aca9a0

use irange for loops (#66234)

r-barnes committed 2 years ago

4734db3b

EmbeddingBackward exclusive_scan thrust->cub (#66566)

zasdfgbnm committed 2 years ago

8159f264

Fix Wextra issues in Half.h (#66643)

r-barnes committed 2 years ago

1b84ad85

Fix torch.cholesky deprecation warning (#66645)

r-barnes committed 2 years ago

995c6642

add BFloat16 operators on CPU: logaddexp, logaddexp2, remainder (#63621)

jiayisunx committed 2 years ago

5f572faf

Fix fx2trt broken unit test (#66696)

Shirong Wu committed 2 years ago

fe05afa3

Newer Older