feat(ci): judge.yml — choix du modèle juge (OpenRouter)#46
Open
benoitvx wants to merge 2 commits into
Open
Conversation
Permet de lancer le LLM-as-a-judge en CI avec un modèle OpenRouter au choix (ex. anthropic/claude-opus-4.8) au lieu du seul défaut. single respecte --judge-provider/--judge-model ; panel accepte un modèle openrouter custom. Le nom du rapport inclut un slug du modèle → pas d'écrasement du run Mistral (gravite-single.md vs gravite-single-anthropic_claude-opus-4.8.md). Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
|
Note Gemini is unable to generate a review for this pull request due to the file types involved not being currently supported. |
Ajoute inputs `judges` (CSV PROVIDER[:MODEL], 2..N juges) et `consensus_min` (nb min de juges confirmant un G3). Permet un vrai consensus N-juges en CI : panel --judge ... --judge ... --judge ... --mode consensus --consensus-min K. Rétro-compatible : panel sans `judges` garde l'ancien comportement. Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Pourquoi
judge.ymlne permettait pas de choisir le modèle juge :singleétait câblé sur Albert/Mistral,panelutilisait le modèle OpenRouter par défaut (sonnet-4.5). Impossible de lancer un juge plus récent (ex.claude-opus-4.8) en CI → on était contraint au local.Changement
Deux inputs
workflow_dispatch:judge_provider(albert | openrouter) — poursingle.judge_model(string) — override du modèle, ex.anthropic/claude-opus-4.8.Câblage :
single→judge --judge-provider <prov> [--judge-model <model>];panel→--judge albert --judge openrouter:<model>si un modèle est fourni. Le nom du rapport inclut un slug du modèle pour ne pas écraser le run Mistral (gravite-single.mdvsgravite-single-anthropic_claude-opus-4.8.md).Note
Le barème G0–G3 est inchangé (on ne fait varier que le modèle juge, pour garder les scores comparables à la baseline).
🤖 Generated with Claude Code