strisunshinewentingwang/pytorch

Pull Requests Commits

Mike Ruberry committed 2 years ago

1f0edf71

Mike Ruberry committed 2 years ago

fbd56d92

Mike Ruberry committed 2 years ago

8ad73215

isnan registration

Mike Ruberry committed 2 years ago

c6d2ee1f

Mike Ruberry committed 2 years ago

9f41f143

adds elementwise opinfos and unary references, extends to out testing

Mike Ruberry committed 2 years ago

238b4e56

Adds python ref consistency test, elementwise unary reference inputs, and formats test files

Mike Ruberry committed 2 years ago

f6bbecf8

[Nvfuser] Add cast support between double and half types

rdspring1 committed 2 years ago

33be4c94

Change mixed precision API

rohan-varma committed 2 years ago

10bf20da

[complex32] support printing the tensor

kshitij12345 committed 2 years ago

e36d25fb

Port functorch decomps over and fix some tests

Chillee committed 2 years ago

fb246140

Provide an auto wrap policy for common transformer models

zhaojuanmao committed 2 years ago

786903ea

Move functorch decompositions to PyTorch

ezyang committed 2 years ago

a3f10ec2

[fix] mul chalf: cpu scalar and cuda tensor case

kshitij12345 committed 2 years ago

1f1d0b30

Nvfuser faster fallback

jjsjann123 committed 2 years ago

100e72f5

Make Opinfo a dataclass to get a more useful repr

francescocastelli committed 2 years ago

201ddafc

Add better error message for FX when using concrete_args

Chillee committed 2 years ago

2540f866

[CUDA Graphs] Fix OOM inside graph capture_begin

fps7806 committed 2 years ago

b0c5fba9

Revert "[Re-landing 68111] Add JIT graph fuser for oneDNN Graph API (Preview4.1)"

pytorchmergebot committed 2 years ago

3dcd67a1

Allow sharding for distributed tests

janeyx99 committed 2 years ago

0708630d

[primTorch] Prototype nvFuser integration and test_prims.py

Mike Ruberry committed 2 years ago

fe1968de

Enable xla test

JackCaoG committed 2 years ago

87a05fa8

[Re-landing 68111] Add JIT graph fuser for oneDNN Graph API (Preview4.1)

chunyuan-w committed 2 years ago

8b11d810

[quant][core][gpu][improvement] Added support for padding quantized cudnn conv2d operator

dzdang committed 2 years ago

8c47e9dc

Add a matching lerp implementation to eager mode. (#1612)

jjsjann123 committed 2 years ago

ac31e5d4

Unsafe Sparse Compressed tensor factory function

pearu committed 2 years ago

ff10e459

Export aten::to("cpu") and aten::to(device="cpu")

Thiago Crepaldi committed 2 years ago

7afe4afd

Fix string repr for nested tensors & run nested tensor tests

jbschlosser committed 2 years ago

fd3cfb68

[GH1] Fix pending/not run conditional and add a way to test new merge_rules

janeyx99 committed 2 years ago

4c014a73

Elias Ellison committed 2 years ago

e5a55af3

Older