ci: fix ci pipeline #90
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Outdated suggestions cannot be applied.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.
Currently, this repo depends on Depot and the build doesn't even work.
In my need to test this repo to fix the failing builds, I stumbled across Blacksmith. Blacksmith is a better fit for RunPod's Containers due to their Layer Caching and pay per use model. These builds are fast and incremental building allows us to move faster than we could previously in improving these models in the future.
Currently, we only automatically build the base model (and we should figure out how to make this scale better!) - but every template in this repo now builds locally. I tried not to change the functionality of any images, just help them build.
It's my hope that the cleanup helps users without very much
docker bake
experience absorb how the repo functions better and in the future actually using our base image to optimize the workflow of images in this repo and outside of it such as https://github.com/runpod/tetra_rp, https://github.com/runpod-workers/worker-template, and users building their own Pod workflows on RunPod.I've also taken the time to drop new builds for CUDA versions less than 12 to match the RunPod fleet, and Python 3.8 due to it no longer being supported by
get-pip.py
.