cache session schema names#2
Open
Goldziher wants to merge 1 commit into
Open
Conversation
`Model::check_schema` was rebuilding a `HashSet<&str>` from `session.inputs()/outputs()` on every inference call. Cache the actual input/output names as `HashSet<String>` once at `Model::new` / `new_from_bytes` (extracted via a new `wrap` helper) and have `check_schema` walk the pipeline's expected iterator against the cached sets — no per-call session-side allocation. The error path still materialises a `HashSet<&str>` for the diagnostic. Also moves schema validation out from under the `inference_lock`: it only reads cached state, so blocking other inferences while we check it is unnecessary.
4 tasks
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Cache the actual session input/output names (
HashSet<String>) once atModel::new/new_from_bytesinstead of rebuilding them on everyinferencecall. Schema validation also moves out from underinference_locksince it only reads cached state.Published as kreuzberg-orp 0.1.1.
Stacks on #1 (
chore/upgrade-ort-rc12). Companion: kreuzberg-gliner-rs PR.Test plan
cargo clippy --all-targets --all-features -- -D warningscargo build --all-targets --all-features