strisunshinewentingwang/pytorch

Pull Requests Commits

Add guidance on the tutorials release proccess (#107871)

Svetlana Karslioglu committed 1 year ago

5c133e91

[ROCm] Add conditions for channels last logic (#107812)

pragupta committed 1 year ago

1ef4bd16

document memory snapshotting (#107660)

zdevito committed 1 year ago

40cbda27

[security] Move s3-html-update workflow into its own environment (#107889)

atalman committed 1 year ago

cd031f13

refactor common code, fix test discovery (#107506)

mlazos committed 1 year ago

6c508e0b

Fix the document of torch.nn.functional.conv2d (#107851)

FFFrog committed 1 year ago

969bf8a0

Fix sym_{sizes,strides} slow path (#107839)

soulitzer committed 1 year ago

f6cce3c4

Fix Inplace tensor update on transpose (#104689)

VRSinghHabana committed 1 year ago

35de780a

Fix aot sequence_nr to reset bwd flag (#107210)

soulitzer committed 1 year ago

3cc5c42a

Revert "[dynamo] Treat monkey patched .forward as dynamic (#107104)"

pytorchmergebot committed 1 year ago

eefce56b

Remove remaining global `set_default_dtype` calls from tests (#107246)

kurtamohler committed 1 year ago

aa8ea1d7

[Easy] use dtype.itemsize in partitions (#107749)

eellison committed 1 year ago

918df101

[dynamo] bugfix - make module setattr more restrictive (#107828)

anijain2305 committed 1 year ago

0156eeb5

Default permissions for torch.hub downloads (#82869)

anibali committed 1 year ago

85b0e03d

make python decomp for native_batch_norm CompositeImplicitAutograd, remove native_batch_norm from core aten opset (#107791)

andrewor14 committed 1 year ago

64d5851b

Fix docstring for shape of `target` for MultiLabelSoftMarginLoss (#107817)

dvorst committed 1 year ago

91a674cc

Check tensors are defined before attempting to access their impl (#106787)

Anstow committed 1 year ago

256fed02

Remove CUTLASS extensions merged upstream (#107612)

alexsamardzic committed 1 year ago

c91d2f5b

remove redundant dynamic_dim (#107815)

avikchaudhuri committed 1 year ago

cf76938f

Ensure optimizer in backward works with 2d parallel (#107748)

rohan-varma committed 1 year ago

8354d32f

[reland][inductor] Adjust dynamic SMEM limit when above default in AOT (#107814)

aakhundov committed 1 year ago

1491bae2

[quant][pt2e] Add support for channel in DerivedQuantizationSpec (#107833)

jerryzh168 committed 1 year ago

16fcb078

[ONNX] Cap opset version at 17 for torch.onnx.export (#107829)

justinchuby committed 1 year ago

38755631

constraint violation error messages (#107790)

avikchaudhuri committed 1 year ago

444875cd

[export] Serialize map correctly (#107837)

angelayi committed 1 year ago

1e71c513

[export] Custom object serialization (#107666)

angelayi committed 1 year ago

1166f9a0

[exportdb] Fix generating docs (#107838)

angelayi committed 1 year ago

6ec2ec84

Revert "inductor: remove conv_bn folding from pre_grad pass (#106686)"

pytorchmergebot committed 1 year ago

2fcda650

Revert "enable conv+bn folding for mixed-dtype when bn has post activation (#107142)"

pytorchmergebot committed 1 year ago

3af04ce0

[vision hash update] update the pinned vision hash (#107831)

pytorchupdatebot committed 1 year ago

6178022a

Newer Older