Decrease thumbnail size for qwen imagery task#415
Merged
Conversation
|
The model we picked will handle image up to 4096^2 pixels (it doesn't have to be square); anything larger gets scaled down by a pre-processing step of some sort, so probably the endpoint would allow even bigger images through. (This does nothing to help the worker's memory use) |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Now that #406 has been merged, I was able to test the behavior in production. My first attempt failed with an OOM kill during the step where we encode the imagery thumbnail (killed before it ever reached out to the inference endpoint). I made a temporary local change to decrease the thumbnail size constant from 4000 to 2000 and tried again, and it worked.
I had increased the thumbnail size to 4000 when we found that the inference endpoint could handle it, but now I find that the limiting factor is the memory of the celery worker, not the memory of the inference endpoint.