Add Mega-Sam image #89
Merged
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
Add the Mega-Sam image built with an update torch-extras image (
torch-extras:es-actions-68fbfd1-nccl-cuda12.4.1-ubuntu22.04-nccl2.25.1-1-torch2.6.0-vision0.21.0-audio2.6.0-abi0
) that updates the old Cuda 11.8 example to ml-containers. Image Page.This image should import and install all dependencies needed up to and including the setup step. We exclude the use of conda as shown in the quickstart in the base repo, as well as the installation of transformers. Instead, follow the To Use section to get to the setup.
To Use
You will want to srun the image with a command like
srun --container-image="ghcr.io#coreweave/ml-containers/cw-mega-sam:rwang.megasam-366a38a" --pty /bin/bash -i
. After that, navigate to/work/mega-sam/base
and runpython setup.py install
.Patch
A cuda function .type() was depreciated and had to be replaced with .type().scalarType() for the build to function properly. This is automatically included in the .patch file, but you may need to manually run this fix if you make changes to the code.