Persian Product Review Sentiment Analysis (mDeBERTa)

Binary sentiment classification (positive/negative) fine-tuned on Persian e-commerce product reviews.

Evaluation Results

Metric Score
Accuracy 87.54%
Macro F1-Score 87.54%
Macro Precision 87.55%
Macro Recall 87.58%

Model Description

This model fine-tunes mDeBERTa on Persian product reviews to classify user sentiment into Positive and Negative categories.

Downloads last month
11
Safetensors
Model size
0.3B params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support