This model is an Arabic language sentiment analysis pretrained model. The model is built on top of the CAMelBERT_msa_sixteenth BERT-based model. We used the HARD dataset of hotels review to fine tune the model. The dataset original labels based on a five-star rating were modified to a 3 label data:

This first prototype was trained on 3 epochs for 1 hours using Colab and a TPU acceleration.

Here are some examples in Arabic to test :

If you have questions or improvement remarks, feel free to contact me on my LinkedIn profile: https://www.linkedin.com/in/yahya-ghrab/