Skip to content

Enabling FP8 all-gather for TE Float8Tensor when using Torch FSDP2 #6402

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

Enabling FP8 all-gather for TE Float8Tensor when using Torch FSDP2 #6402

Triggered via pull request December 5, 2024 01:39
Status Failure
Total duration 2m 13s
Artifacts

lint.yml

on: pull_request
PyTorch C++
19s
PyTorch C++
PyTorch Python
2m 3s
PyTorch Python
JAX C++
22s
JAX C++
JAX Python
23s
JAX Python
PaddlePaddle C++
23s
PaddlePaddle C++
PaddlePaddle Python
51s
PaddlePaddle Python
Fit to window
Zoom out
Zoom in

Annotations

1 error and 6 warnings
PyTorch Python
Process completed with exit code 16.
PyTorch C++
ubuntu-latest pipelines will use ubuntu-24.04 soon. For more details, see https://github.com/actions/runner-images/issues/10636
JAX C++
ubuntu-latest pipelines will use ubuntu-24.04 soon. For more details, see https://github.com/actions/runner-images/issues/10636
JAX Python
ubuntu-latest pipelines will use ubuntu-24.04 soon. For more details, see https://github.com/actions/runner-images/issues/10636
PaddlePaddle C++
ubuntu-latest pipelines will use ubuntu-24.04 soon. For more details, see https://github.com/actions/runner-images/issues/10636
PaddlePaddle Python
ubuntu-latest pipelines will use ubuntu-24.04 soon. For more details, see https://github.com/actions/runner-images/issues/10636
PyTorch Python
ubuntu-latest pipelines will use ubuntu-24.04 soon. For more details, see https://github.com/actions/runner-images/issues/10636