I have used the FLAIR library for text classification. I was amazed to find it produces high accuracy with pre-trained embedding on different datasets.
I was wondering how this library implements its text classification (sentiment analysis). Looking into the code of the text classification I saw that flair uses RNN but also uses distilBERT in some way.
Could anyone elaborate on exactly how flair does sentiment analysis on a task?
Thank you!
flair lib: https://github.com/flairNLP/flair/ flair
Text Classification: https://github.com/flairNLP/flair/blob/master/flair/models/text_classification_model.py
flair_sentiment = flair.models.TextClassifier.load('en-sentiment')
