ai21labs / sense-bert Goto Github PK

Hi，

As I know, word sense is case-sensitive.

Have you ever trained the cased version? Could you please share the checkpoint？

Thank you very much for sharing the code for your excellent paper.
Pardon me for asking this newbie question: how to align the tokens in the input sentence with the supersenses outputted from the model?
For example, the words in the sentence "I went to the store to buy some groceries." do not appear to be aligned with the following senses

['noun.person']
['verb.communication']
['verb.social']
['verb.communication']
['noun.artifact']
['noun.artifact']
['verb.communication']
['verb.cognition']
['noun.artifact']
['noun.artifact']
['adv.all']
['adv.all']

as printed using the following code:

for i, id_ in enumerate(input_ids[0]):
  print(sensebert_model.tokenizer.convert_ids_to_senses([np.argmax(supersense_logits[0][i])]))

Could you please provide some example code for how to do this properly? Thanks a lot in advance!

Hey,

A very interesting paper, thanks a lot for this contribution!
Do you have plans to upload a model on huggingface?

Best,
Max

Hi! Can you please release a version compatible with newer TensorFlow?
Thanks!

Hi. Thank you very much for sharing the code for loading your models. I am trying to use the SenseBert model to get the sense of some words in tweets and for that, I want to fine-tune your model for a simple sequence classification that classifies tweets into offensive and non-offensive. Could you share the code that would be able to do this or tell me how I could do this?

Thanks in advance!

Hello. Thank you very much for sharing your code. I'm a student and I want to reproduce your results on WiC competition. As I understood, in the inference you get two supersenses for the word in different contexts, after that you compare these two supersenses and decide if they are in the same meaning or not. The question is, do you finetune the model, if so, how is it done?
As I see it, you have to take supersense logits, after that compare them, may be with CosineSimilarity and get the loss.

guys are u planning to release sensebert training code and dataset , if so when @yoel-zeldes @oriram