strisunshinewentingwang/pytorch

Pull Requests Commits

[functorch] Some more test annotations

zou3519 committed 2 years ago

01164988

[functorch] re-annotated some tests

zou3519 committed 2 years ago

e86d8ebf

[functorch] Refactor DynamicLayer so the logic for each transform is separate (pytorch/functorch#756)

zou3519 committed 2 years ago

3bf44893

[functorch] Add xfails for nansum, nanmean

zou3519 committed 2 years ago

8a6491c8

[functorch] Only generate vmap plumbing for operations in the allowlist (pytorch/functorch#758)

zou3519 committed 2 years ago

2788574f

[functorch] nll loss decomp (pytorch/functorch#755)

Samantha Andow committed 2 years ago

60214759

[functorch] L1 loss decomposition (pytorch/functorch#753)

Samantha Andow committed 2 years ago

b2b1a00d

[functorch] fix binary_cross_entropy_backward decomposition and fix _tensor_str

Chillee committed 2 years ago

ac9a47bf

[functorch] xfail binary cross entropy decomposition on cuda, bfloat16

samdow committed 2 years ago

fd05b533

[functorch] DynamicLayerBack refactor (pytorch/functorch#754)

zou3519 committed 2 years ago

8aaea274

[functorch] Fix pad failures

zou3519 committed 2 years ago

851b4e72

[functorch] Refactor dynamicLayerFrontFallback (pytorch/functorch#751)

zou3519 committed 2 years ago

927e2a6b

[functorch] updated some decompositions and cleaned some stuff

Chillee committed 2 years ago

9eb10d03

[functorch] Add l1 loss forward decompositions

Chillee committed 2 years ago

151b1f4a

[functorch] Add l1 forward decomposition

Chillee committed 2 years ago

da3eb084

[functorch] fix issue

Chillee committed 2 years ago

1204b542

[functorch] Add embedding decomposition

Chillee committed 2 years ago

8625aaee

[functorch] missed one unexpected success

samdow committed 2 years ago

669b0113

[functorch] fix unexpected successes (pytorch/functorch#752)

Samantha Andow committed 2 years ago

57765636

[functorch] Fix some of the block_diag failures

zou3519 committed 2 years ago

ab280e5a

[functorch] add diag_embed batch rule

zou3519 committed 2 years ago

c71dfa6e

[functorch] DynamicLayer uses TransformType instead of DispatchKey to identify transforms (pytorch/functorch#748)

zou3519 committed 2 years ago

99279f46

[functorch] Add expand_copy and masked_select_backward to reenable tests from upstream breakages (pytorch/functorch#736)

Samantha Andow committed 2 years ago

4b2ded61

[functorch] Added params_require_grad arg to make_functional* (pytorch/functorch#701)

vfdev-5 committed 2 years ago

34b20941

[functorch] Generalize get_decompositions handling for OpOverloadPackets (pytorch/functorch#743)

ezyang committed 2 years ago

d5ec0bfb

[functorch] Fixed flake8 error on CI (pytorch/functorch#744)

vfdev-5 committed 2 years ago

47c75a87

[functorch] functionalize(): properly handle global return tensors (pytorch/functorch#735)

bdhirsh committed 2 years ago

c5c4f746

[functorch] skip flaky(?) test

Chillee committed 2 years ago

e7293ee5

[functorch] fix random stuff

Chillee committed 2 years ago

9c65dd57

[functorch] Add a decomposition and remove op_analysis from .gitignore

Chillee committed 2 years ago

46142ab4

Newer Older