AventIQ-AI/sentiment_analysis_product_review_sentiment

🧠 SentimentClassifier-BERT-TweetEval

A BERT-based sentiment analysis model fine-tuned on the TweetEval dataset. It predicts the sentiment of a text as Positive, Neutral, or Negative with confidence scores. This model is useful for classifying product feedback, user reviews, or social media posts.

✨ Model Highlights

• 📌 Based on bert-base-uncased (by Google)
• 🔍 Fine-tuned on the Sentiment Subtask from TweetEval
• ⚡ Supports prediction of 3 classes: Negative, Neutral, Positive
• 💾 Available in both full and quantized versions for inference

🧠 Intended Uses

• Product release analysis (e-commerce, apps)

🚫 Limitations

• Not optimized for other languages than English
• May not generalize well to domains very different from Twitter/product reviews
• Can confuse sarcasm or irony
• Performance may degrade on long texts (>128 tokens, due to truncation)

🏋️‍♂️ Training Details

• Base Model: bert-base-uncased
• Dataset: TweetEval: Sentiment Subtask
• Framework: PyTorch with 🤗 Transformers
• Epochs: 5
• Batch Size: 8
• Max Length: 128 tokens
• Optimizer: AdamW
• Loss: CrossEntropyLoss with class balancing
• Device: Trained on NVIDIA CUDA-enabled GPU

📊 Evaluation Metrics

Metric	Score
Accuracy	0.99
F1-macro	0.98

Replace the above with actual scores after evaluation.

🔎 Label Mapping

Label ID	Sentiment
0	Negative
1	Neutral
2	Positive

🚀 Usage

Load the Model

from transformers import BertTokenizer, BertForSequenceClassification
import torch
import torch.nn.functional as F

model_name = "AventIQ-AI/sentiment_analysis_product_review_sentiment"
tokenizer = BertTokenizer.from_pretrained(model_name)
model = BertForSequenceClassification.from_pretrained(model_name)
model.eval()

def predict(text):
    inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True, max_length=128)
    with torch.no_grad():
        outputs = model(**inputs)
        probs = F.softmax(outputs.logits, dim=1)
        pred = torch.argmax(probs, dim=1).item()
        label_map = {0: "Negative", 1: "Neutral", 2: "Positive"}
        return f"Sentiment: {label_map[pred]} (Confidence: {probs[0][pred]:.2f})"

# Test predictions
print("\nTest Predictions:")
print(predict("We're thrilled to announce our latest update, packed with new features and performance improvements!"))

Quantization

Post-training quantization was applied using PyTorch's built-in quantization framework to reduce the model size and improve inference efficiency.

Repository Structure

.
├── model/               # Contains the quantized model files
├── tokenizer_config/    # Tokenizer configuration and vocabulary files
├── model.safensors/     # Fine Tuned Model
├── README.md            # Model documentation

Limitations

• The model may not generalize well to domains outside the fine-tuning dataset.
• Quantization may result in minor accuracy degradation compared to full-precision models.

Contributing

Contributions are welcome! Feel free to open an issue or submit a pull request if you have suggestions or improvements.