Dixie is capped (max_inflight 2, queue_depth 8, queue_wait 20s) and never queued on llama-server. Streaming chat, thinking off, 45s fallback to clean. Digest and chunk-embedding caches are content-addressed SQLite (Honcho context is not cached). Recover uses the guarded Fetcher. guardian-star is refused in code.
956 B
956 B
How to contribute
We definitely welcome patches and contribution to this project!
Tips
Commits must be formatted according to the Conventional Commits Specification.
Always try to include a test case! If it is not possible or not necessary, please explain why in the pull request description.
Releasing
Commits that would precipitate a SemVer change, as described in the Conventional
Commits Specification, will trigger release-please
to create a release candidate pull request. Once submitted, release-please
will create a release.
For tips on how to work with release-please, see its documentation.
Legal requirements
In order to protect both you and ourselves, you will need to sign the Contributor License Agreement.
You may have already signed it for other Google projects.