ask some files in process_data.py

[Z-yolo]

Thank you very much for your work open source, but we are more curious about some files in your process_data, such as train/valid/test_parsed.json, are these files made by you or where did you get them, I wonder if it is convenient for you to tell us, thanks a lot! with open('./erc_data/meld/train_parsed.json') as f: parsed_train = json.load(f) with open('./erc_data/meld/valid_parsed.json') as f: parsed_valid = json.load(f) with open('./erc_data/meld/test_parsed.json') as f: parsed_test = json.load(f)

[circle-hit]

Thanks for your attention! We upload the parsed data files and the way to get them. For any other question, feel free to contact me.

[Z-yolo]

Thank you very much for your offer, and good luck with your research!

[yeefighting]

There are two questions I would like to ask, the first one: regarding the discourse_graph, discourse_speaker_mask, inter_speaker_graph, intra_speaker_graph, intra_relative_distance, inter_relative_distance, mentioned in meld/meld_features_roberta_discourse.pkl, how are these derived; and the second: how the implicit detachment mentioned in the paper was derived？

[circle-hit]

There are two questions I would like to ask, the first one: regarding the discourse_graph, discourse_speaker_mask, inter_speaker_graph, intra_speaker_graph, intra_relative_distance, inter_relative_distance, mentioned in meld/meld_features_roberta_discourse.pkl, how are these derived; and the second: how the implicit detachment mentioned in the paper was derived？

Q1: These can be obtained by running process_data.py. Q2: Implicit detachment is derived from the attention martrix of MHA (line 97&98 in models/mucdn.py)

[yeefighting]

Thank you very much for your offer and have a nice life！