Dockerfile for ROCm #311
Draft
Dockerfile for ROCm #311
+37
−0
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
apparently the build caching works a lot better with a plain venv, so apart from a few seconds refreshing repos this step only takes time on first install anyways.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
This is a first Draft of a usable Dockerfile for AMD ROCm cards.
Based on my testing it currently only works with the
raw-pngbranch, probably due to the newer dependencies (9.0.7 on AMD is also broken with micromamba).The flash_attn handling is also not that pretty, but I'm not sure how to improve it. Feedback is especially welcome here.
I'll also have to do some performance testing to see whether there's any negative impact, and if newer ROCm versions (6.1.2) or even potentially mismatched versions (6.0 on 6.2 or 6.1 on 6.2) improve things.
There might also STILL be some memory leaks during model load, but using Docker and venv compard to micromamba makes bisecting easier.