Fix reading modules.json for Dense modules in local models
#738
Conversation
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
As that might imply that the user originally provided a local path rather than a Hugging Face Hub ID, meaning that the `dense_paths` variable won't be filled, meaning that we need to read those from `modules.json` Note that this is just a premature quick solution, ideally this should be handled within `backends/src/lib.rs` rather than directly within the `CandleBackend` as otherwise we end up duplicating a lot of unnecessary code
4 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
What does this PR do?
This PR fixes the missing
Densemodule handling for local models, that was preventing from runningtext-embeddings-inferencewith a local copy of a model as e.g.google/embeddinggemma-300m, given that it was only handled when the provided--model-idwas a Hugging Face Hub model ID, hence the artifacts where downloaded on runtime.Note
This is just a tentative solution to patch the aforementioned issue, but the handling of the
modules.jsonas well as other modules asNormalize,Pooling,TransformersorStaticEmbeddingshould be handled in an unified way.Fixes #732
Before submitting
instasnapshots?Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.