Spaces:

autoevaluate

/

model-evaluator

like 172

Running

Hello,

I'm trying to use the Model Evaluator UI to evaluate my roberta finetuned model on the validation portion of the squad_v2 dataset. When I finish completing the fields and click 'Submit', I encounter a KeyError involving the key 'eval_info' arising from line 21 of evaluation.py:

metadata = dataset_info.cardData["eval_info"]

Is this an issue with the dataset repository card, or something I'm doing incorrectly?