Irony_Detection_roberta_bertweet

URL: https://github.com/Martinaeht/Irony_Detection_roberta_bertweet

Description: Hyperparameter tuning of roberta and bertweet on irony detection dataset cardiffnlp/tweet_eval

Methods

This repository focuses on the hyperparameter tuning of two pre-trained models, RoBERTa and BERTweet, for the task of irony detection. The models are fine-tuned on the cardiffnlp/tweet_eval dataset, which is designed for tweet-based sentiment and irony detection.

The primary method involves experimenting with different hyperparameters for both RoBERTa and BERTweet, evaluating their performance on the irony detection task.

Results

The repository does not provide detailed results or metrics directly in the description. However, it is implied that the hyperparameter tuning process aims to improve the performance of RoBERTa and BERTweet on the irony detection dataset.

Dataset

The repository uses the cardiffnlp/tweet_eval dataset, which contains annotated tweet data specifically designed for evaluating sentiment and irony detection models. The dataset is available for download from the Hugging Face dataset repository.

Dataset Reference: cardiffnlp/tweet_eval

Dataset Link: cardiffnlp/tweet_eval