search

zerohd4869 / SACL

The repository for ACL 2023 paper "Supervised Adversarial Contrastive Learning for Emotion Recognition in Conversations", and SemEval@ACL 2023 paper "UCAS-IIE-NLP at SemEval-2023 Task 12: Enhancing Generalization of Multilingual BERT for Low-resource Sentiment Analysis"
MIT License
29 stars 6 forks source link

The settings of hyperparameters for training #6

Closed SkylarShow closed 4 months ago

SkylarShow commented 6 months ago

Hello! I would like to know if I need to modify some hyperparameter values if I want to train a new model by removing the contrastive learning part. My paper is related to your project, so I would really appreciate it if you could answer this question!

zerohd4869 commented 4 months ago

Hello! I would like to know if I need to modify some hyperparameter values if I want to train a new model by removing the contrastive learning part. My paper is related to your project, so I would really appreciate it if you could answer this question!

If you are using the same model backbone as our SACL-LSTM, you can refer to the sh script presented in the "Quick Start" section of this repo for the hyperparameter search range. Thank you!