BERT-ASTD Balanced

Arabic version bert model fine tuned on ASTD dataset balanced version to identify twitter sentiments in Arabic language MSA dialect .

Data

The model were fine-tuned on ~1330 tweet in Arabic language.

Results

class	precision	recall	f1-score	Support
0	0.9328	0.9398	0.9363	133
1	0.9394	0.9323	0.9358	133
Accuracy			0.9361	266

How to use

You can use these models by installing torch or tensorflow and Huggingface library transformers. And you can use it directly by initializing it like this:

from transformers import AutoModelForSequenceClassification, AutoTokenizer

model_name="mofawzy/BERT-ASTD"
model = AutoModelForSequenceClassification.from_pretrained(model_name,num_labels=2)
tokenizer = AutoTokenizer.from_pretrained(model_name)