- Download the 4 models and the MultiLabelBinarizer file
mlb.pklfrom this link. Use gdown if you are trying to download to a remote instance via terminal because wget or curl is insufficient to download large files from google drive. - Place the models inside of the
modelsdirectory and unzip.
In auto_eval_utterances.py, the function evaluate_utterances(utterances, csv_name) should be called to evaluate your utterances or chats. The parameter utterances is a list of strings, and the parameter csv_name is a string name for the output csv that contains the autoamtic evaluations.
Now you can just call evaluate_utterances(utterances, csv_name) to evaluate a list of utterances with the 4 models you downloaded above and see the results inside of the evaluations folder.