Align default argument values with evaluation repository #377
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
Update args_parser.py and benchmark-specific run_infer.py files to use default values that match the evaluation repository (OpenHands/evaluation) eval-job/values.yaml configuration. Shared defaults updated in args_parser.py: - workspace: 'docker' -> 'remote' - max-iterations: 100 -> 500 - critic: 'pass' -> 'finish_with_patch' Benchmark-specific overrides using parser.set_defaults(): - gaia: dataset='gaia-benchmark/GAIA' - swtbench: dataset='eth-sri/SWT-bench_Verified_bm25_27k_zsp' - commit0: max_attempts=1, max_retries=1 (in addition to existing dataset) Also updated AGENTS.md to document the default values alignment pattern. Co-authored-by: openhands <openhands@all-hands.dev>
…hmultimodal - swebench: Add explicit set_defaults(dataset, split) for consistency with other benchmarks, even though values match global defaults - swebenchmultimodal: Update comment to match the pattern used in other benchmarks Co-authored-by: openhands <openhands@all-hands.dev>
Each benchmark now sets its own dataset default via set_defaults(), so no global default is needed. Co-authored-by: openhands <openhands@all-hands.dev>
All benchmarks in the evaluation repository use .llm_config/runtime.json as the LLM config path, so use this as the default. Co-authored-by: openhands <openhands@all-hands.dev>
This reverts commit c34d730.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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
This PR aligns the default argument values in the benchmarks repository with the values used in the evaluation repository (
OpenHands/evaluation)eval-job/values.yamlconfiguration.Changes
Shared defaults updated in
args_parser.py:--workspacedockerremote--max-iterations100500--criticpassfinish_with_patchBenchmark-specific overrides using
parser.set_defaults():gaiadataset="gaia-benchmark/GAIA"swtbenchdataset="eth-sri/SWT-bench_Verified_bm25_27k_zsp"commit0max_attempts=1, max_retries=1(in addition to existing dataset)swebenchprinceton-nlp/SWE-bench_Verified)swebenchmultimodaldataset,split="dev")Documentation
AGENTS.mdwith documentation about the default values alignment patternBenefits
set_defaults()Testing
py_compilesyntax validation@simonrosenberg can click here to continue refining the PR