strisunshinewentingwang/pytorch

Pull Requests Commits

remove unused all_generator_source from generate_code (#75868)

mikey dagitses committed 2 years ago

f010b15d

move version_h to shared build structure in Buck (#75964)

mikey dagitses committed 2 years ago

8b1cf8ed

Support 8 inputs, UT Passing

SherlockNoMad committed 2 years ago

17fe8116

Add NVFuser Python Frontend

Kevin Stephano committed 2 years ago

b17b2b1c

use c10::variant

SherlockNoMad committed 2 years ago

aa623a95

Revert "[JIT] Allow registering Decompositions"

pytorchmergebot committed 2 years ago

2d72cb33

Reapply black for torchgen, this time with lint to fix!

ezyang committed 2 years ago

c2ae0b01

[quant][docs] Fix formatting for quantization.rst (#76223)

jerryzh168 committed 2 years ago

30342f6b

[JIT] Allow registering Decompositions

Elias Ellison committed 2 years ago

d9f0774f

Fix memory use after free: buffer need to be attached to the module. (#76350)

qihqi committed 2 years ago

6d0f2be3

Fix bug where __getstate__ of DDP looks for self._replicated_tensor_module

pritamdamania87 committed 2 years ago

b26df43f

E2E SymInt example narrow_copy

Krovatkin committed 2 years ago

bb60cac2

Add black linter

ezyang committed 2 years ago

d286197e

deploy: add dummy metadata for builtin packages (#76211)

d4l3k committed 2 years ago

bbc6fcd7

Fix issue in sparce_coo_tensor only supporting CUDA device.

chengjunlu committed 2 years ago

a9deda54

supports various num inputs

SherlockNoMad committed 2 years ago

ec865118

Update torch::lazy::BackendDevice to have a new default ordinal (#76264)

wonjoolee95 committed 2 years ago

f9d07ae6

[lint] correctly display annotations for all severities

suo committed 2 years ago

1e111895

[reland] use scatter in shard_tensor API (#75991)

wanchaol committed 2 years ago

1f3aa535

Fix forward-mode AD formula for binary_cross_entropy_with_logits

zou3519 committed 2 years ago

9cb2871f

patching 11.1 ptxas issue

jjsjann123 committed 2 years ago

e48b29b1

functionalization bugfix: using owning type when unwrapping tensors

bdhirsh committed 2 years ago

640ce6bc

[NVFuser] multithreading nvfuser test

davidberard98 committed 2 years ago

f36d348f

Avoid compiling ts backend binidngs in fbcode build

wconstab committed 2 years ago

a4ce1136

[POC] fix static init issue with JIT container types

bdhirsh committed 2 years ago

40f3e850

remove _is_foreach_op codegen special cases, clean up mutable return type checks

bdhirsh committed 2 years ago

74e93f72

functionalization: add native fill() op

bdhirsh committed 2 years ago

ea5209c9

functionalization: add a copy() native function

bdhirsh committed 2 years ago

5da76acd

functionalization: add support for zero_()

bdhirsh committed 2 years ago

7d44b367

[complex32] mul

kshitij12345 committed 2 years ago

c8fd1bbc

Newer Older