strisunshinewentingwang/pytorch

Pull Requests Commits

[BE] Use ValueError for input.dim check in torch.nn.modules (#105127)

isdanni committed 1 year ago

b33d63d9

[foreach][RMSprop] Minimize intermediates=2 to decrease peak memory (#105161)

janeyx99 committed 1 year ago

cd152299

[foreach][ASGD] Minimize intermediates=1 to decrease peak memory (#105146)

janeyx99 committed 1 year ago

219cf2a1

Serialize empty pytree cases (#105159)

angelayi committed 1 year ago

3a7d77f7

move kineto submodule commit (#105144)

briancoutinho committed 1 year ago

6c10edcb

Dynamo tensor aliasing guards, dedup graphargs (#104921)

voznesenskym committed 1 year ago

485cad4a

Reland: Make `torch.empty*` deterministic by filling with NaN or max int (#104995)

kurtamohler committed 1 year ago

f987d11f

[ONNX] Fix UnsupportedFxNodesAnalysis after onnx dispatcher changes (#105156)

BowenBao committed 1 year ago

42530c17

[ONNX] Apply param_manipulation.py from onnx-script to op validation and dispatcher (#104679)

titaiwangms committed 1 year ago

15c1e44d

Add semi-structured sparse conversions (#103830)

alexsamardzic committed 1 year ago

fc2f87b2

Revert "[export] Allow optional call-spec (#105041)"

pytorchmergebot committed 1 year ago

15478a50

[Dockerfile] Add `make triton` to the `build` target (#105114)

yncxcw committed 1 year ago

df3a64fb

Use plain power operator in Adam/Adamw when capturing (#104254)

albanD committed 1 year ago

ef05c5f2

[export] Allow optional call-spec (#105041)

angelayi committed 1 year ago

194fe1d1

Fix typo (#105130)

janEbert committed 1 year ago

b06a4263

SDPA: frontend for BSR masks (#104042)

nikitaved committed 1 year ago

44c8515d

[dynamo] Simulate torch function enablement state (#105091)

mlazos committed 1 year ago

05eea20e

Switch automatic_dynamic_shapes to True by default in fbcode (#104883)

ezyang committed 1 year ago

87cf51cc

Revert "[inductor] Register an op for mm_plus_mm (#104835)" (#105150)

clee2000 committed 1 year ago

c36dca7b

Revert "[inductor] fix a custom_op test problem (#104972)" (#105149)

clee2000 committed 1 year ago

91c64f55

Perform value range analysis with rationals when possible (#105137)

lezcano committed 1 year ago

d1fedad0

Update mypy to 1.4.1 (#91983)

malfet committed 1 year ago

634659e2

enable channels last for reflection padding on CPU (#102518)

mingfeima committed 1 year ago

f73757d5

Revert "[PyTorch TB] Write raw tensor as tensor_proto (#104908)"

pytorchmergebot committed 1 year ago

d35137cc

Add some more files to ciflow/inductor (#105112)

ezyang committed 1 year ago

e1502c0c

[inductor] fix incorrect strides in copy() decomp, fix hf_LongFormer + hf_BigBird errors (#100115)

bdhirsh committed 1 year ago

c6b9c31a

Optimize scatter_add/scatter_reduce in BFloat16/Half data type in CPU backend (#103427)

yanbing-j committed 1 year ago

053654b9

[dynamo] Maintainable code - Move decorators in a separate file (#105070)

anijain2305 committed 1 year ago

735e6ae8

Update attention.cpp to remove warning about preferring torch.bool type (#103362)

mikekgfb committed 1 year ago

4a0d773a

Transmute refined SymInt into int (#104828)

ezyang committed 1 year ago

0f322a30

Newer Older