Lint

Enabling FP8 all-gather for TE Float8Tensor when using Torch FSDP2 #6441

Sign in to view logs

Run time

Learn about OS pricing on GitHub Actions

Job	Run time
JAX C++	20s
JAX Python	22s
PaddlePaddle C++	19s
PaddlePaddle Python	56s
PyTorch C++	24s
PyTorch Python	2m 6s
	4m 27s