Keep K3 suite selection and report-schema scoring in bash, merge K3/vision dataset_args into dpv4 yamls, and pin EvalScope at 735d920ee911 with local patches. Co-authored-by: Cursor <cursoragent@cursor.com>
578 B
578 B
扩展评测基准
本章节介绍需要额外安装依赖包或进行特殊配置的评测基准。这些基准通常针对特定领域或任务,提供了更专业的评测能力。
使用这些基准前,请先按照各基准文档中的说明安装相应的依赖包和完成必要的环境配置。
:::{toctree} :maxdepth: 1
terminal_bench.md miniwob.md skillsbench.md toolathlon.md gaia.md wide_search.md deepsearchqa.md swe_bench.md swe_bench_pro.md tau_bench.md tau2_bench.md tau3_bench.md bfcl_v3.md bfcl_v4.md needle_haystack.md toolbench.md longwriter.md :::