Would it be possible to include information relative to the original filenames of the evaluation dataset in the predictions_eval.json file? I.e. to associate a given list of mentions in the predictions file to the original filename (through a key or something)?
Hi
Thank you for your work!
Would it be possible to include information relative to the original filenames of the evaluation dataset in the predictions_eval.json file? I.e. to associate a given list of mentions in the predictions file to the original filename (through a key or something)?