[BETA] Adaptive Router#43
Open
jeann2013 wants to merge 1 commit into
Open
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
[BETA] Adaptive Router
What it does: A virtual model (auto_router/adaptive_router) that classifies each request into one of seven types, chooses the best available real-world model by balancing learned quality versus cost, and learns from satisfaction signals in subsequent turns. Adapted to /v1/messages and /v1/responses (this project does not have /v1/chat/completions).
Added:
Config: model_info.adaptive_router_preferences, litellm_params.adaptive_router_config, with full validation.
Engine (src/sdk/adaptive_router/): Regex classifier, scoring with cold-start prior + online bandit (maximum 200 samples), conversation correlation without session ID (hash of the first message).
Persistence in Postgres (0003_adaptive_router_cells.sql), with proper degradation without a database.
Header x-litellm-adaptive-router-model + endpoint GET /adaptive_router/{router_name}/state.
docs/adaptive_router.md + example in config.yaml.example.
Testing:
51 unit tests + 4 HTTP integration tests (wiremock).
fmt, clippy -D warnings, check_code_size.py — clean.
Manually verified with a real server: Real Postgres + Real Anthropic API. Confirmed end-to-end: classification → model selection → header → correlation between two requests → "thanks!" signal → bandit update (0.95→1.0, samples 0→1) — seen in both the HTTP response and the Postgres query.