Add /run-evals comment command#1948
Open
adibarra wants to merge 2 commits into
Open
Conversation
…orkflow plumbing) Adds the workflow-side plumbing so a PR comment can launch an eval-only run of a single recipe (no perf sweep): /run-evals <eval> <config-key> [conc] [master-config] - run-evals.yml: new issue_comment command (mirrors pr-comment-sweep.yml auth/ SHA-pin/reply); maps <eval> -> framework+task, infers nvidia/amd master config from the config-key HW token, builds 'test-config ... --evals-only', calls e2e-tests.yml with eval-framework/eval-task. - e2e-tests.yml: new eval-framework/eval-task inputs, threaded into the eval jobs. - benchmark-tmpl.yml / benchmark-multinode-tmpl.yml: matching inputs -> env (EVAL_FRAMEWORK / EVAL_TASKS_DIR). Inert by default (eval-framework defaults to lm-eval). The framework-dispatch code (run_eval EVAL_FRAMEWORK override + run_swebench_eval + swebench task/scorer) lives with the swebench PR and is checked out from the commented PR's head at runtime, so this can merge to main independently.
Collaborator
|
@adibarra isnt there eval-only tag that does this? |
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
Collaborator
Author
|
Going for a slightly different functionality here compared to the existing one |
- run-evals.yml: 'Reply with run link' now uses always() + branches on the parse step outcome, so a bad eval name / missing config-key / unrecognized HW token gets a helpful PR reply instead of silently doing nothing. - run-evals.yml: document the gpqa_diamond/swebench aliases in the header + both error messages (were only in the case statement). - e2e-tests/benchmark-tmpl/benchmark-multinode-tmpl: align the eval-framework input description across all three (drop the misleading 'recipe default' — the consumer falls back to a hardcoded lm-eval).
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
2 participants
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.
Adds the workflow-side plumbing so a PR comment can launch an eval-only run of a single recipe (no perf sweep):
/run-evals [conc] [master-config]
Inert by default (eval-framework defaults to lm-eval). The framework-dispatch code (run_eval EVAL_FRAMEWORK override + run_swebench_eval + swebench task/scorer) lives with the swebench PR and is checked out from the commented PR's head at runtime, so this can merge to main independently.