search

aviclu / CDLM

49 stars 10 forks source link

How to use the pretrained model for inference #10

Open jama1017 opened 1 week ago

jama1017 commented 1 week ago

Hi,

Thanks so much for the amazing work. I am quite new to the huggingface api and I wonder if you can kindly let me know how I can just use the pretrained CDLM model for event coreference resolution?

aviclu commented 1 week ago

Hi,

Here are the weights for entity coreference: https://drive.google.com/drive/folders/179JFsUyIiPaqspKyo7rBxUFzcB79xHzq?usp=sharing

and for event coreference: https://drive.google.com/drive/folders/1bVSI3VPAvcIixrclkZyTHDXi7bamg93b?usp=sharing

jama1017 commented 1 week ago

Hi!

Thanks so much for your reply. I was looking for a bit more specific guidance on how to write the code.

Specfically, I can run the following as indicated on README:

from transformers import AutoTokenizer, AutoModel
# load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained('biu-nlp/cdlm')
model = AutoModel.from_pretrained('biu-nlp/cdlm')

But I wonder how I can use this model on a new example document with events identified, not just to evaluate on the preloaded datasets?