CYBER WEEK - EXTRA SAVINGS EVENT
TRY A FREE LESSON

Multi Text Classification

I am building a text classifier using DistillBERT from Huggingface which classify input text into Geography, Environment and Science , How could I increase the accuracy for first two labels as both are always predicted as Science because of the similarity ? What are the tweaks should I do? or any other thing. Please Guide me.

Hi, @Lavkushgupt

Did you tune the model to your data?
How much marked-up data do you have?

If yes and it didn’t help. Check model repository, maybe there is a pre-tuned model which was trained on scientific texts.

Regards, Max