Benhamdane/Sentiment-analysis-darija

--- language: ar tags: - sentiment-analysis - darija - arabic license: apache-2.0 datasets: - custom metrics: - accuracy - precision - recall - f1 --- # Sentiment Analysis for Darija (Arabic Dialect) This repository hosts a **Sentiment Analysis model for Darija** (Moroccan Arabic dialect), built using **BERT**. The model is fine-tuned to classify text into two categories: **positive** and **negative** sentiment. It is designed to facilitate sentiment analysis in applications involving Darija text data, such as social media analysis, customer feedback, or market research. --- ## Model Details - **Base Model**: SI2M-Lab/DarijaBERT - **Task**: Sentiment Classification (Binary) - **Architecture**: BERT with a custom classification head and dropout regularization (0.3 dropout rate). - **Fine-Tuning Data**: Dataset of labeled Darija text samples (positive and negative). - **Max Sequence Length**: 128 tokens --- ## How to Use ### Load the Model and Tokenizer To use this model for sentiment analysis, you can load it using the Transformers library: ```python from transformers import AutoTokenizer, AutoModelForSequenceClassification # Load the tokenizer and model tokenizer = AutoTokenizer.from_pretrained("BenhamdaneNawfal/sentiment-analysis-darija") model = AutoModelForSequenceClassification.from_pretrained("BenhamdaneNawfal/sentiment-analysis-darija") # Example text test_text = "هذا المنتج رائع جدا" # Tokenize the text inputs = tokenizer(test_text, return_tensors="pt", truncation=True, padding=True, max_length=128) # Get model predictions outputs = model(**inputs) logits = outputs.logits predicted_class = logits.argmax().item() print(f"Predicted class: {predicted_class}") ``` ### Output Classes - **0**: Negative - **1**: Positive --- ## Fine-Tuning Process The model was fine-tuned using the following: - **Dataset**: A dataset of Darija text labeled for sentiment. - **Loss Function**: …