strisunshinewentingwang/pytorch

Pull Requests Commits

[CUDA][CUBLAS] Fix BF16 reduced precision reduction note in Numerical accuracy docs (#101884)

lkct committed 1 year ago

807d8115

[export] Prototype on serialization schema. (#101899)

zhxchen17 committed 1 year ago

351c2ea2

[MPS] Fix embedding cache key (#101857)

qqaatw committed 1 year ago

330c9073

Partially fix shape mismatch in vision_maskrcnn (#101477)

ezyang committed 1 year ago

22ca1a11

[vision hash update] update the pinned vision hash (#101938)

pytorchupdatebot committed 1 year ago

9e8da7fb

[T153220354] Fix header inclusions in c10 (#1541) (#101846)

q10 committed 1 year ago

66a2600b

Prevent pattern matches across mutation ops in inductor pre-grad FX passes (#101144)

williamwen42 committed 1 year ago

dde6d561

disableing quantizing gradient in 8bw (#101739)

kwanghoon-meta committed 1 year ago

13640bf9

[ONNX] Bump onnx submodule to release 1.14.0 (#101809)

titaiwangms committed 1 year ago

f0dc41a7

group constraints by arg (#101815)

avikchaudhuri committed 1 year ago

03de1580

Disallow module forward input mutation in aot_export (#101834)

tugsbayasgalan committed 1 year ago

b5ee34e5

[inductor] Refactor RNG operators (#100064)

jansel committed 1 year ago

0c6f409c

[dynamo] Minor refactor to use is_allowed to decide inlining of NNModule methods (#101910)

anijain2305 committed 1 year ago

8b2a9f81

[vision hash update] update the pinned vision hash (#101917)

pytorchupdatebot committed 1 year ago

2886b3e6

inductor: fix name 'inf' is not defined issue when calling external_call function (#101865)

XiaobingSuper committed 1 year ago

bb62a373

inductor: fix bfloat16 store complier issue (#101856)

XiaobingSuper committed 1 year ago

350f0cd7

[accuracy minifier] cast copied model rather than update the original model (#101901)

shunting314 committed 1 year ago

029c6a99

[easy] refactor signature flattening transform (#101886)

avikchaudhuri committed 1 year ago

73e887b5

[dynamo] Bugfix for unspecialized nn module variable (#101859)

anijain2305 committed 1 year ago

7a17e9d0

[inductor] Test indirect indexing asserts with dimension of size 1 (#101811)

peterbell10 committed 1 year ago

48346a46

[inductor] Implement magic methods on IR values (#101076)

peterbell10 committed 1 year ago

89bd5d3d

[quant][pt2e] Introduce QuantizationAnnotation API (#101708)

jerryzh168 committed 1 year ago

15495f2d

[codemod][3.10][NamedTuple] Use typing_extensions to get NamedTuple Generics (#101830)

fried committed 1 year ago

03f50fcc

Fix finding existing Needs label comments (#101889)

kit1980 committed 1 year ago

b07e97c0

[pt2] Turn off lazy reinit when cuda graph is on (#101848)

xw285cornell committed 1 year ago

c8fd1cfa

[Profiler] Workaround CUPTI Lazy Reinit and CUDA Graphs crash in CUDA 11 (#101879)

aaronenyeshi committed 1 year ago

fa7ad77a

Dynamic Shape Doc (#101885)

msaroufim committed 1 year ago

3666ca9d

Fake Tensor Docs (#101882)

msaroufim committed 1 year ago

ff5b9428

Add Logging Doc to compile index (#101888)

msaroufim committed 1 year ago

581d13a0

Revert "separate out dynamo .requires_grad and .is_grad_enabled guards (#100570)"

pytorchmergebot committed 1 year ago

7f3fed12

Newer Older