Replies: 1 comment 1 reply
-
I think the issue probably comes from here. if model_name in [
"vinai/bertweet-base",
"vinai/bertweet-covid19-base-cased",
"vinai/bertweet-covid19-base-uncased",
]:
self.tokenizer = tokenizer_class.from_pretrained(
model_name,
do_lower_case=self.args.do_lower_case,
normalization=True,
**kwargs,
)
else:
self.tokenizer = tokenizer_class.from_pretrained(
model_name, do_lower_case=self.args.do_lower_case, **kwargs
) A quick fix would be to just add the name of the model you want to use here. I will change this logic later to check for strings matching any Bertweet model. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hello,
I am experimenting with NER, more concretely, the Bertweet model (I saw that simpletransformers supports this model), and I noticed that vinai/bertweet-large does not work. Can you please tell me whether this a bug or this model is not supported? I have tested the other betrtweet models: vinai/bertweet-base, vinai/bertweet-covid19-base-cased, and vinai/bertweet-covid19-base-uncased, and they work properly.
Thank you in advance and I would like to use this opportunity to thank you for the great job you are doing: simpletransformers is an awesome library!
Best regards,
Igor
Beta Was this translation helpful? Give feedback.
All reactions