strisunshinewentingwang/pytorch

Pull Requests Commits

Add size check before calling stack_.at(dict_pos) in unpickler.cpp (#94300)

m4drat committed 1 year ago

794e3971

Use yaml.SafeLoader instead of legacy yaml.Loader (#100443)

kit1980 committed 1 year ago

ab65bac3

Add error_inputs_sparse method to OpInfo (#100389)

pearu committed 1 year ago

02a0fb8d

Replace master with main in links and docs/conf.py (#100176)

Svetlana Karslioglu committed 1 year ago

d425da8b

Support expandable_segments:True in fbcode for caching allocator

zdevito committed 1 year ago

0aac2446

Mount /etc/hosts into container (#100475)

malfet committed 1 year ago

99ded8bb

Revert "[functorch] test for compiling functorch transforms (#100151)"

pytorchmergebot committed 1 year ago

af92fc1c

Initial version of Dynamo capture for HigherOrderOperator (#99988)

zou3519 committed 1 year ago

4c99f9cd

Refactor OutputGraph (#99987)

zou3519 committed 1 year ago

984a2397

Revert "[pytorch] Accelerate indexing_backward_kernel with duplicates (#99441)"

pytorchmergebot committed 1 year ago

1114673c

[inductor] Remove redundant model copy when running with cpp_wrapper (#100275)

desertfire committed 1 year ago

ec3c8abb

[export] Migrate internal verifier to subclass export/verifier

angelayi committed 1 year ago

Verified af62d098

[pt2] add `SymInt` support for `tensordot` and `inner` (#100356)

nkaretnikov committed 1 year ago

41361538

[distributed][sharded_tensor] Move local_shards check from ShardedTensorBase to ShardedTensor (#100197)

IvanKobzarev committed 1 year ago

4582ceb2

Make backend_accuracy_fails suppress errors in same_two_models (#100324)

ezyang committed 1 year ago

8556cf20

Run minifier tests same process when possible (#100416)

ezyang committed 1 year ago

054a254b

Prevent Triton from getting eagerly imported when importing torch._inductor (#100374)

ezyang committed 1 year ago

f093ee17

Speed up minifier tests by applying some configs that speed things up. (#100387)

ezyang committed 1 year ago

74cc3771

Simplify minifier testing by incorporating fault injection in prod code (#100357)

ezyang committed 1 year ago

0a479d9b

Do not use pickle to output config entries in repro scripts (#100354)

ezyang committed 1 year ago

17be6538

Manually resolve download.pytorch.org to IPv4 (#100436)

huydhn committed 1 year ago

0093df78

[dynamo] Graph break on a list referencing self (#100296)

anijain2305 committed 1 year ago

52a36a98

Only skip in torch inductor test

houseroad committed 1 year ago

Verified 090ec55f

Use a stable ordering for saved values in functorch.default_partition (#100111)

jamesr66a committed 1 year ago

d5169e71

[functorch] test for compiling functorch transforms (#100151)

kshitij12345 committed 1 year ago

ea5f6d73

[inductor] Prevent reusing aliased buffers if aliases still have uses (#100332)

ngimel committed 1 year ago

ff297223

[vision hash update] update the pinned vision hash (#100437)

pytorchmergebot committed 1 year ago

3fd46e1f

Add --baseline option to benchmark runners (#100266)

jansel committed 1 year ago

fdc853b1

Delete @property support at module level, it is unused (#100353)

ezyang committed 1 year ago

c6c92583

Disable densenet121 as it is flaky (#100371)

ezyang committed 1 year ago

e918fd18

Newer Older