MZ

Hi professor, I find that there is no No.60 dialogue in data/MELD/train.csv, can you check it? Best wishes, Mian

Hi, In https://github.com/yxuansu/MAGIC/blob/main/story_generation/language_model/utlis.py#L266, you say you only use the generated text to calculate similarity. But in the paper, the MAGIC scores are calculated based on the prefix and the generated...