strisunshinewentingwang/pytorch

Pull Requests Commits

[ONNX] Add supported ops into test_fx_op_consistency - 1st batch (#100265)

titaiwangms committed 1 year ago

843ead13

[SPMD] add FQN argument to Override.replacement (#100473)

mrshenli committed 1 year ago

2ebb48ff

Use `maybe_unused` attr in VariableType (#100498)

crcrpar committed 1 year ago

6cc01583

Revert "Initial version of Dynamo capture for HigherOrderOperator (#99988)"

pytorchmergebot committed 1 year ago

58f796ff

Stop loading functorch._C unless torchdim is needed (#100491)

zou3519 committed 1 year ago

b2d703e2

[fix] torch_compile_debug don't log with 0 (#100462)

kshitij12345 committed 1 year ago

8b64dee5

[Dynamo] Skip TB Background_Matting model eager accuracy check because of non deterministic (#100513)

yanboliang committed 1 year ago

896eb1db

Retry resolving download.pytorch.org with Google DNS (#100509)

huydhn committed 1 year ago

9e2808aa

[PT2E][Quant] Refactor quantizer and qnnpack qantizer code to support dqlinear config (#99399)

kimishpatel committed 1 year ago

771a9deb

Relax after_aot restriction on no buffers, serialize small constants (#100472)

ezyang committed 1 year ago

1bbca4fb

Refactor minifier tests to be more compact (#100471)

ezyang committed 1 year ago

2089a9bd

Make hash_storage work with size 0/1 storage (#100467)

ezyang committed 1 year ago

409fc7a4

Allow discontiguous NestedTensors to empty_like (#98383)

drisspg committed 1 year ago

4b9ba3fa

Run periodic jobs only twice a day on weekends (#100489)

kit1980 committed 1 year ago

419387f6

OpInfo: specifying sparse sample input function implies the corresponding layout support (#100392)

pearu committed 1 year ago

6b2ecb12

Fix sum OpInfo for sparse sample inputs and assert coverage for sparse-enabled operators (#100391)

pearu committed 1 year ago

3ae0e23b

Move tracked nn_modules from OutputGraph to TracingContext (#100457)

voznesenskym committed 1 year ago

ffcbd1c2

Remove special casing for stride/size setup for guards (#100456)

voznesenskym committed 1 year ago

2439090b

Avoid using einsum for torch.cat DTensor propogation (#100251)

mrshenli committed 1 year ago

9439cb0e

[ONNX] Add RemoveConstantInputStep to adapt torch inputs to ONNX inputs (#100252)

titaiwangms committed 1 year ago

d23dbfff

[inductor] Change the default value of layout (#100254)

desertfire committed 1 year ago

6b5f5000

Revert "[WIP] enable cuda graphs support for flash attention with dropout (#100196)"

pytorchmergebot committed 1 year ago

c3aa59c8

Fix hosts update for binary build (#100507)

malfet committed 1 year ago

dc4a2531

Remove ci: sev label and details from ci-sev.md tempalte (#100504)

kit1980 committed 1 year ago

e4ad67f9

Revert "[inductor] Cleanup strip_last_size logic (#100305)"

pytorchmergebot committed 1 year ago

34e90b8d

[PT2E][Quant] Fix but in quant spec of symmetric static quant (#99398)

kimishpatel committed 1 year ago

8ec0a939

Handle trailing masked column behavior for nested tensor (#100113)

Michael Gschwind committed 1 year ago

8430430e

[dashboard] higher tolerance for AlbertForQuestionAnswering (#100277)

anijain2305 committed 1 year ago

0acfe2ce

[inductor] Cleanup strip_last_size logic (#100305)

jansel committed 1 year ago

de7793d5

[WIP] enable cuda graphs support for flash attention with dropout (#100196)

Natalia Gimelshein committed 1 year ago

32615618

Newer Older