Replies: 2 comments 4 replies
-
参考这里的代码来配置具体的Evaluator和后处理方式: 给你的datasets加上一个eval_cfg,比如: datasets = [{"path": "xxx/opencompass-test-qa.jsonl", "data_type": "qa", "infer_method": "gen", "eval_cfg": dict(evaluator=dict(type=AccwithDetailsEvaluator),pred_postprocessor=dict(type=first_capital_postprocess))}] 如果你需要自定义metric,则在对应的位置自己写一个Evaluator即可 |
Beta Was this translation helpful? Give feedback.
3 replies
-
我也遇到了一样的问题,并且尝试配置了eval_cfg,但似乎没用起到作用,请问您解决了吗 |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
使用opencompass自定义数据集,如何配置评测指标?自定义qa类型数据集,通过openApi的方式进行评估,发现评估结果中使用的是 accuracy。具体如下:
评估配置内容如下:
opencompass-test-qa.jsonl内容如下:
每个任务的推理结果(predictions):
通过以上结果来看,结果不该为0,因此如何配置评测指标,即metric的值?
Beta Was this translation helpful? Give feedback.
All reactions