strisunshinewentingwang/pytorch

Pull Requests Commits

Rewrite docs about why we guard on dynamic dims (#120566)

ezyang committed 208 days ago

fd3cf88f

[export] Change runtime asserts to using assert_scalar (#119608)

angelayi committed 208 days ago

75920425

Enable fake tensor caching in fbcode by default (#118555)

masnesral committed 208 days ago

2fb32a5f

[dynamo] Function => FunctionCtx for placeholder obj (#120577)

jansel committed 208 days ago

ee01d080

[inductor] Optimize welford reduction (#120330)

peterbell10 committed 208 days ago

7eb7ac81

Numbers based TD (#119901)

clee2000 committed 208 days ago

c39bbd6d

Add torch._print to dynamo trace_rules (#120533)

angelayi committed 208 days ago

86063b4d

Revert "Add meta device support to sparse compressed tensors (#120498)"

pytorchmergebot committed 208 days ago

8a32a078

make GPT2ForSequenceClassification pass inference accuracy check (#120537)

shunting314 committed 208 days ago

b381a437

Fix a bug where nn.functional._AllGather.backward produces wrong gradients (#120582)

yifuwang committed 208 days ago

f4cf25bb

Add resnet50/mobilenet_v2_quantized_qat in into deterministic_algorithms exclusive list (#120384)

leslie-fang-intel committed 208 days ago

c617e7b4

[dynamo][guards-cpp-refactor] NO_HASATTR guard (#120469)

anijain2305 committed 208 days ago

a299db29

[inductor][cpp] disable masked load for non-fp data types (#120558)

jgong5 committed 208 days ago

1c7b0e7c

[executorch hash update] update the pinned executorch hash (#120264)

pytorchupdatebot committed 208 days ago

ea20885d

[dynamo] Reland 120147 - - Use EQUALS_MATCH guard for mod.training (#120578)

anijain2305 committed 208 days ago

c18623b7

Add SparsePrivateUse1 in backend_to_string, layout_from_backend and check_base_legacy_new. (#119263)

Shan19900305 committed 208 days ago

685d862c

[dynamo][guards-cpp-refactor] DICT_VERSION guard (#120416)

anijain2305 committed 208 days ago

4328e772

[dynamo][guards-cpp-refactor] DictGuardManager (#120359)

anijain2305 committed 208 days ago

c269e48a

[dynamo][guards-cpp-refactor] WEAKREF_ALIVE guard (#120344)

anijain2305 committed 208 days ago

775a4388

Add meta device support to sparse compressed tensors (#120498)

pearu committed 209 days ago

5d71ba68

[dynamo][refactor] Move some helper functions to global scope (#120426)

anijain2305 committed 209 days ago

834c7a1d

Fix default world_size when running on 1 or 0 GPU (#119372)

tinglvv committed 209 days ago

5c7b761f

[Clang-tidy header][19/N] Enable clang-tidy on torch/csrc/autograd/profiler_legacy.* (#120552)

cyyever committed 209 days ago

81f0b2c1

[dynamo] support group=None when rewriting collectives (#120118)

yifuwang committed 209 days ago

298c686d

Fix compiler check (#120492)

xuhancn committed 209 days ago

3e382456

Don't use size on TensorVariable when doing out resize test (#120567)

ezyang committed 209 days ago

0f20cc1e

add distributed checkpoint support for custom device (#120201)

ChuBoning committed 210 days ago

54c1cf8d

Add profiling for backward (#120540)

mlazos committed 210 days ago

56203fc4

[dynamo] add stronger test for dynamo memory leaks (#120459)

williamwen42 committed 210 days ago

a17979fa

[ET-VK] Move graph runtime from PT directory to ET directory (#120528)

SS-JIA committed 210 days ago

a62d9184

Newer Older