Issues training the model #4
Description
Hi there,
First of all great paper!
Currently I'm trying to train the model locally so that I can properly reproduce the paper.
I'm running on a few issues when training, specifically:
- When running the training file, an error pops up at the utils.py file where, the QA_action_reason file only has 3 examples, whereas your code requires for 15 - which prompts the exception in line 48. I assume the file which i got from the official Ads Dataset website is not the right now to use for training.
- When i change the number of examples in the utils.py to accept anything under 15, which in this case is 3 examples, the code fails after loading persuasion strategies, with an error saying that the model accepts 512 tokens, but got 609 tokens.
I'm assuming the above errors to be related to inappropriate files being used. If that is the case, Would you be able to share these files, as used in your project?
If the files seem to be alright based on my description, do you have any advice on running it?
kind regards,