search

Alibaba-NLP / ACE

[ACL-IJCNLP 2021] Automated Concatenation of Embeddings for Structured Prediction
Other
300 stars 44 forks source link

About config for tweebank dataset #3

Closed hary00078 closed 3 years ago

hary00078 commented 3 years ago

Hello First of all, I read the your paper published well. I was very impressed. I applied uploaded code to the CoNLL 2003 data set. I confirmed the performance suggested in the paper. This time, I want to confirm performance about POS with Tweebank data set. But I can't confirm proper performance. So shall you send me the yaml file for POS? Thanks for reading.

wangxinyu0922 commented 3 years ago

Hi, I have uploaded the config file. You can find the config file in config/. Before train the model, you need to finetune BERT, M-BERT and XLM-R on the tweebank at first. You can find the guide for finetuning transformer-based models at: https://github.com/Alibaba-NLP/ACE#optional-fine-tune-transformer-based-embeddings

hary00078 commented 3 years ago

Thank you. I'll do it right away