search

raspberryice / gen-arg

Code for paper "Document-Level Argument Extraction by Conditional Generation". NAACL 21'
MIT License
119 stars 30 forks source link

data process #19

Closed Hou-jing closed 2 years ago

Hou-jing commented 2 years ago

the data you provide is right?? when I use preprocessed_KAIROS this file ,I don't find the following items which used in dataloader: 'input_token_ids': input_token_ids, 'input_attn_mask': input_attn_mask, 'tgt_token_ids': tgt_token_ids, 'tgt_attn_mask': tgt_attn_mask, 'doc_key': doc_keys, image

raspberryice commented 2 years ago

Hi Hou-jing, The preprocessed_KAIROS directory is supposed to be created by the KAIROS_data_module.py file and it should contain the tokenized results. It is not where you should put your source data. See the prepare_data method for details. image

raspberryice commented 2 years ago

If you are not getting this preprocessed_KAIROS automatically, you can first run this prepare_data separately.

Hou-jing commented 2 years ago

Thanks, I understand.

Hou-jing commented 2 years ago

Sorry, I come across another question,I can not find Keyword-Based Trigger Extraction Model which mentioned in your paper. Did you put it in on this repository？

raspberryice commented 2 years ago

Please check the tapkey branch.

Hou-jing commented 2 years ago

Thank you, I find it