Disable "vocab_article_source.json"

[Aichiniuroumian]

Hi~@Nicozwy Thank you so much for your amazing work, I have already star your work. In the process of replicating your code, I also encountered an error that the file "vocab_article_source.json" was not found. I have tried the following method you gave me: 1, You can set ' 'vocab_article_source=None" in "tran_xxx.py "because the file "vocab_article_source.json" is no longer used. 2, You can simply create a new file with the name "vocab_article_source.json" because it is not used in the codes. But I have encountered a loading error in the "vocab_article_source.json" file. Do I need to change anything other than the above solution? Looking forward to your reply, have a good day!

[Nicozwy]

Setting ' 'vocab_article_source=None" in "train_xxx.py" can solve this problem. Try this https://drive.google.com/file/d/1QpqiK_R9PGG9NesTYqC8aLJhMZ0CYQ00/view?usp=sharing

[Aichiniuroumian]

Setting ' 'vocab_article_source=None" in "train_xxx.py" can solve this problem. Try this https://drive.google.com/file/d/1QpqiK_R9PGG9NesTYqC8aLJhMZ0CYQ00/view?usp=sharing

Thank you very much for your timely reply! I tried the "train_exp_fc5_LIAR_RAW2-backup.py" file you provided, but it still failed to load. I have noticed that the 'vocab_article_source.json' file is used in the "source_url = pjoin(data_url,filename)" in "reader5.py". May I ask how to comment it here? I'm very sorry to bother you, thank you again ~🙂

[Nicozwy]

You should try to disable the loading function because ''vocab_article_source.json" is no longer used.
By the way, you may check what you have changed or directly use debugging mode.

[Aichiniuroumian]

You should try to disable the loading function because ''vocab_article_source.json" is no longer used. By the way, you may check what you have changed or directly use debugging mode.

Thanks for your timely reply, I changed the two lines in reader5.py to the following, and it can run through. # lm_ids_dict['report_domains'] = gen_domain_id(raw_data_list[6], device=_device) lm_ids_dict['report_domains_ids'] = [] I would also like to ask you a question. What does Ac, p, r and maf1 at the front of the training log mean? Is there any difference between them and eA, eP, eR and emaf1 at the back? 2023-09-15 21:12:11,103 - INFO: Ep[1]-Batch[2000/5033] Loss: 117.0914 [1.71:1.13:180.00], Ac: 0.0000, p: 0.0000, r: 0.0000, maf1: 0.000000|| eA: 0.8462, eP: 0.5031, eR: 0.5085, emaf1: 0.470164 Thank you again for your kind help!😊

[JasonGuo1]

Thank you for your kind help and discuss.However, the training process doesn't seem quite right from the figure above. Did I make any mistakes？