Task Adapters

Hinglish Sentiment

This dataset was released as part of SemEval 2020, Task 9 on Sentiment Analysis in Code Mixed Social Media (Twitter) text. It tags positive, neutral and negative sentiment. There are 17,000 tweets, of which 14,000 are marked for training.

Website

sentiment/hinglish-twitter-sentiment@nirantk bert-base-multilingual-uncased

1 version Architecture: pfeiffer Head:

Adapter for Hinglish Sentiment Analysis, based on SemEval 2020 Task 9