Paul Zhang

tolist() support for FunctionalTensor

1

Summary: Support tolist() for FunctionalTensor for KJT in torch.export Reviewed By: ezyang Differential Revision: D53731064

Introduce pt2 ir export unit test

2

Summary: Introduce unit testing torch.export Differential Revision: D53426083

Introduce Training benchmarks

3

Differential Revision: D53481057

CPU Unit Test CI to 15 minutes

1

Differential Revision: D52631694

Allow Missing Labels for MTML models

3

Differential Revision: D50273422

Summary: D56021085 - This fixed copying FP parameters to meta device when sharding model on meta Weird thing is though that FP parameters are not sparse parameters, they are dense....

fix model split boundary

2

Summary: As title Differential Revision: D56148751

batch_size * lengths for export

6

Differential Revision: D55602984

cpu ci timeout

1

Differential Revision: D54756659

test tolist() on KJT with torch export

1

Differential Revision: D54490404