Add endpoint for emptying the model cache #7602
Merged
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
github-actions
bot
added
api
python
RyanJDick
force-pushed
the
ryan/empty-model-cache
branch
from
3c2a65a to
3613ee3
Compare
RyanJDick
requested review from
psychedelicious,
blessedcoolant,
maryhipp,
hipsterusername,
lstein and
brandonrising
as code owners
hipsterusername
approved these changes
Jan 30, 2025
…k to the ModelCache to make it thread-safe.
hipsterusername
force-pushed
the
ryan/empty-model-cache
branch
from
eb5bac9 to
24a078b
Compare
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Outdated suggestions cannot be applied.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.
Summary
This PR adds a backend endpoint that can be used to empty the model cache (to free RAM / VRAM). 'Locked' models that are actively being used will not be dropped.
Locking was added to
ModelCacheto make it thread-safe. But, I suspect that it might be possible to trigger #7513 with an ill-timed request to empty the cache (in the middle of graph execution). Even if this does happen, the current graph execution would fail, but subsequent graph executions should recover smoothly.QA Instructions
I tested that the new endpoint successfully empties the model cache with
curl -X POST 127.0.0.1:9090/api/v2/models/empty_model_cache.I tried to break things by hitting the endpoint throughout graph execution. Nothing broke, which is a good sign. Of course, this is far from testing the thread interactions thoroughly.
Merge Plan
No special instructions.
Checklist
What's Newcopy (if doing a release after this PR)