Closed Eta0 closed 1 year ago
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5719570265
Image: ghcr.io/coreweave/ml-containers/torch:es-flash-attn-2-41a6446-base-cuda12.1.1-torch2.0.1-vision0.15.2-audio2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5719570265
Image: ghcr.io/coreweave/ml-containers/torch:es-flash-attn-2-41a6446-base-cuda12.0.1-torch2.0.1-vision0.15.2-audio2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5719570265
Image: ghcr.io/coreweave/ml-containers/torch:es-flash-attn-2-41a6446-base-cuda11.8.0-torch2.0.1-vision0.15.2-audio2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5719570264
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-41a6446-nccl-cuda11.8.0-nccl2.16.2-1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn1.0.9
Did this successfully build earlier?
@wbrown On a testing container, yes. The code is valid. On our CI, it's only finishing most of the time. Seems to be resource-constrained. Requested some more powerful runners from @todie (related: #25) in lieu of dropping the concurrency even lower.
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5719570264
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-41a6446-nccl-cuda11.8.0-nccl2.16.2-1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5719570264
Image: ghcr.io/coreweave/ml-containers/torch:es-flash-attn-2-41a6446-nccl-cuda12.0.1-nccl2.18.3-1-torch2.0.1-vision0.15.2-audio2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5719570264
Image: ghcr.io/coreweave/ml-containers/torch:es-flash-attn-2-41a6446-nccl-cuda12.1.1-nccl2.18.3-1-torch2.0.1-vision0.15.2-audio2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5719570264
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-41a6446-nccl-cuda12.1.1-nccl2.18.3-1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5719570264
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-41a6446-nccl-cuda12.1.1-nccl2.18.3-1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn1.0.9
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5721928378
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-750fc34-nccl-cuda12.0.1-nccl2.18.3-1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5721928378
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-750fc34-nccl-cuda12.0.1-nccl2.18.3-1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn1.0.9
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5721944894
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-750fc34-nccl-cuda11.8.0-nccl2.16.2-1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5721944894
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-750fc34-nccl-cuda11.8.0-nccl2.16.2-1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn1.0.9
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5722249778
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-5fbc6bb-nccl-cuda12.1.1-nccl2.18.3-1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn1.0.9
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5722249778
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-5fbc6bb-nccl-cuda12.1.1-nccl2.18.3-1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5722319221
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-5fbc6bb-base-cuda12.1.1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5722319221
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-5fbc6bb-base-cuda12.1.1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn1.0.9
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5722324333
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-5fbc6bb-base-cuda12.0.1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5722324333
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-5fbc6bb-base-cuda12.0.1-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn1.0.9
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5722326374
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-5fbc6bb-base-cuda11.8.0-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn2.0.2
@Eta0 Build complete, success: https://github.com/coreweave/ml-containers/actions/runs/5722326374
Image: ghcr.io/coreweave/ml-containers/torch-extras:es-flash-attn-2-5fbc6bb-base-cuda11.8.0-torch2.0.1-vision0.15.2-audio2.0.2-flash_attn1.0.9
This change updates
flash-attention
in thetorch-extras
container to version 2.0.2.The
.github/workflows/torch-extras.yml
workflow now publishesflash-attention
1.x and 2.x variants of thetorch-extras
container, distinguished by the suffixes-flash_attn<VERSION>
. Right now the variants are:-flash-attn2.0.2
-flash-attn1.0.9