Skip to content
PR GPU tests

Bumping flash attention version to 2.6.3 and adding option for softcap in attention and lm_head logits. #7756

Sign in to view logs

Bumping flash attention version to 2.6.3 and adding option for softcap in attention and lm_head logits.

Bumping flash attention version to 2.6.3 and adding option for softcap in attention and lm_head logits. #7756

Triggered via pull request July 24, 2024 20:58
@ShashankMosaicMLShashankMosaicML
synchronize #1374
Status Success
Total duration 10m 31s
Artifacts –

pr-gpu.yaml

on: pull_request_target
Matrix: pytest-gpu-1
Matrix: pytest-gpu-2
Matrix: pytest-gpu-4
Fit to window
Zoom out
Zoom in

Annotations

3 warnings
gpu-2.3.1-4 / pytest-gpu
The following actions uses Node.js version which is deprecated and will be forced to run on node20: actions/checkout@v3, actions/setup-python@v4, actions/cache@v3. For more info: https://github.blog/changelog/2024-03-07-github-actions-all-actions-will-run-on-node20-instead-of-node16-by-default/
gpu-2.3.1-2 / pytest-gpu
The following actions uses Node.js version which is deprecated and will be forced to run on node20: actions/checkout@v3, actions/setup-python@v4, actions/cache@v3. For more info: https://github.blog/changelog/2024-03-07-github-actions-all-actions-will-run-on-node20-instead-of-node16-by-default/
gpu-2.3.1-1 / pytest-gpu
The following actions uses Node.js version which is deprecated and will be forced to run on node20: actions/checkout@v3, actions/setup-python@v4, actions/cache@v3. For more info: https://github.blog/changelog/2024-03-07-github-actions-all-actions-will-run-on-node20-instead-of-node16-by-default/