strisunshinewentingwang/pytorch

Pull Requests Commits

[Inductor] Add Int8 data type into Inductor CPP backend vectorized code generation (#119179)

jgong5 committed 227 days ago

a050d146

Avoid COW materialize in pooling, batch linalg, upsample, softmax ops (#119503)

kurtamohler committed 227 days ago

5918622d

Avoid COW materialization for TensorInfo with const type (#119502)

kurtamohler committed 227 days ago

53deddd6

Avoid COW materialization for TensorAccessors with const type (#119501)

kurtamohler committed 227 days ago

fba5b7f7

Clarifying windows cosine behaviour in the documentation (#119444)

jmarintur committed 227 days ago

fa071a2e

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

masnesral committed 227 days ago

0f2fbbff

[BE] Update requests to 2.31.0 (#119516)

malfet committed 227 days ago

2cdf9b76

Revert "Add FakeTensor support to torch._utils._rebuild_tensor (#108186)"

pytorchmergebot committed 227 days ago

458e83b5

Add Debug Utility To Generate Inputs for AOT Graphs (#119409)

eellison committed 227 days ago

930b60f5

Don't log canonicalized expressions (#119471)

lezcano committed 227 days ago

2d474e17

Revert "Fix jagged NT softmax semantics (#119459)"

pytorchmergebot committed 227 days ago

8994f236

[inductor] Add split scan kernel (#117992)

peterbell10 committed 227 days ago

88429a80

[inductor] Refactor triton range_tree handling (#117991)

peterbell10 committed 227 days ago

01edb8a5

Update chunk_dtensor to support HYBRID_SHARD (#119481)

mvpatel2000 committed 227 days ago

6efda849

Disable tests that use bfloat 16 for SM < 80 (#118449)

tinglvv committed 227 days ago

454abb6b

[Dynamo] Support kwargs for lazy module (#119445)

yoyoyocmu committed 227 days ago

915f9db0

Update setup tools to 65.5.1 (#119456)

malfet committed 227 days ago

45c4a0ce

Revert "Add lowering for logcumsumexp (#118753)"

pytorchmergebot committed 227 days ago

a8d1645f

[DeviceIndex] Use DeviceIndex instead of int in CUDA wrappers (#119142)

cyyever committed 227 days ago

560c92c3

[ROCm] enable hipsolver backend for linalg.eigh (#115177)

jeffdaily committed 227 days ago

e98dbae0

[export] allow user input mutation in aot_export (#119356)

suo committed 227 days ago

0f12c0af

[aot_inductor] replace TORCH_CHECK with AOTI_CHECK in the generate cpp code (#119220)

chenyang78 committed 227 days ago

9f8ade04

Update logging.cpp for explicit chrono import (#119469)

qlzh727 committed 227 days ago

71e772f8

Windows Dynamo Error Removal CI Check (#115969)

gs-olive committed 227 days ago

45e7af58

[export] Remove torch._export.export (#119095)

angelayi committed 227 days ago

0827510f

[dtensor] switch softmax backward ops to OpStrategy (#119255)

tianyu-l committed 227 days ago

a7754b2b

Fixed an issue where nn.Linear would cause an internal int underflow … (#119221)

tringwald committed 227 days ago

d9a1b258

s/print/warn in arch choice in cpp extension (#119463)

msaroufim committed 227 days ago

7fd6b1c5

[BE] Add dtypesIfMPS to ModuleInfo enabling float16 tests for MPS and remove all skipIfMPS for float64 (#119039)

mikaylagawarecki committed 227 days ago

db1a4dcb

[Inductor] Setting kernel launch and exit callbacks for inductor generated triton kernels (#119450)

Jokeren committed 227 days ago

4e93b00b

Newer Older