strisunshinewentingwang/pytorch

Pull Requests Commits

Dynamo - config gated torch.distributed allow, exclusion for special leaf funcs (#110894)

voznesenskym committed 1 year ago

395d0eae

Use CUDA image for lintrunner (#110502)

cyyever committed 1 year ago

49914635

[Dist][ez][nit] Formatted nccl version string in startup (#111076)

rohan-varma committed 1 year ago

d8ad0ba5

[DDP] Make _ReplicateState inherit from _State and make replicate eagerly initialized (#109647)

fegin committed 1 year ago

b35279df

Move export.constrain_as_* to torch._constrain_as_* (#110757)

tugsbayasgalan committed 1 year ago

5614023f

Revert "Move export.constrain_as_* to torch._constrain_as_* (#110757)"

pytorchmergebot committed 1 year ago

6ce3a380

[vision hash update] update the pinned vision hash (#111098)

pytorchupdatebot committed 1 year ago

f0e7a910

Allow specifiying inputs as GradientEdge in autograd APIs (#110867)

albanD committed 1 year ago

5e8be63e

[export] Fix graph signature data model to list of specs. (#111017)

zhxchen17 committed 1 year ago

33b69509

[device mesh] only check when world size > num_devices per host (#111091)

wanchaol committed 1 year ago

097defb1

Use torch._check for cat error checking (#111035)

ezyang committed 1 year ago

9316c8b4

Revert 107846 and 109695 (#111099)

kit1980 committed 1 year ago

6dca81c0

update tensor-like to check instance for torch function impl (#111087)

mlazos committed 1 year ago

07f0f383

[TP] Validate TP mesh dim for 2D composition (#111001)

wz337 committed 1 year ago

8e32e62f

Bump xla_base version tag to v1.1 (#109757)

yeounoh committed 1 year ago

80ea8784

[quant][pt2e][be] Move xnnpack quantizer tests to separate file (#111004)

jerryzh168 committed 1 year ago

e0ddc3ff

Linear Quantize (#110581)

shubhraprakash1 committed 1 year ago

8f8d8a0b

Add `torch.unravel_index` (#110580)

kurtamohler committed 1 year ago

5292a92e

[aotinductor] Fail models temporarily (#111100)

angelayi committed 1 year ago

577e3dff

[2/N] Dynamo supports skip by function & removes skipfiles circular import (#110835)

yanboliang committed 1 year ago

986ad3bf

[2/N] Enable Wunused-result, Wunused-variable and Wmissing-braces in torch targets (#110836)

cyyever committed 1 year ago

a6b452df

Fix typo under torch/_functorch directory (#111067)

kiszk committed 1 year ago

6d7744ca

torch.compile DTensor E2E (#105236)

bdhirsh committed 1 year ago

4d29b402

Add CUTLASS-based support for mixed dtypes matrix multiplication (#110981)

cpuhrsch committed 1 year ago

3553eb9b

Fix `functional::smooth_l1_loss` signatures to not override `beta` (#109798)

kurtamohler committed 1 year ago

0f924cde

[reland2] Update custom Function preserve torch function when inputs … (#110895)

soulitzer committed 1 year ago

73f4c1a4

fix ShardedTensor.gather when shard is empty (#110962)

winter-yi committed 1 year ago

52e76a30

AOTAutograd: Go down inference path if no outputs require grad (#111011)

bdhirsh committed 1 year ago

ded5ee75

[export] Get export APIs ready for PTC (reland) (#111030)

angelayi committed 1 year ago

6d8e0c4b

Fix Android publish step with lite interpreter (#111071)

huydhn committed 1 year ago

20a73661

Newer Older