This model is a fully fine-tuned version of distilbert-base-uncased for 5-class star rating prediction (1-5 stars) on Singapore digital bank reviews. All ~67M parameters were updated during training.

🎯 Model Performance

Metric	Score
Test Accuracy	0.4130
Precision (Macro)	0.4461
Recall (Macro)	0.4130
F1-Score (Macro)	0.3268
MAE (Mean Absolute Error)	1.0217 stars

🚀 Quick Start

Loading the Model

python

from transformers import AutoModelForSequenceClassification, AutoTokenizer
import torch

# Load model and tokenizer
model = AutoModelForSequenceClassification.from_pretrained("ajiayi/bert-full-finetuned-singapore-digi-banks")
tokenizer = AutoTokenizer.from_pretrained("ajiayi/bert-full-finetuned-singapore-digi-banks")

# Make predictions
text = "Love this bank! Best interest rates in Singapore."
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=128)
outputs = model(**inputs)
predicted_stars = outputs.logits.argmax(dim=1).item() + 1  # Convert 0-4 to 1-5
print(f"Predicted rating: {predicted_stars} stars")

📊 Model Details

Model Architecture

Base Model: distilbert-base-uncased (66M parameters)
Fine-tuning Method: Full supervised fine-tuning (all parameters updated)
Task: Multi-class text classification (5 classes)
Classes: 1-star, 2-star, 3-star, 4-star, 5-star

Model Size

Total Parameters: 66,957,317
Model Size: ~250MB
Training Time: 1.99 minutes

🎓 Training Details

Training Data

Domain: Singapore digital bank reviews (GXS, Trust Bank, MariBank, etc.)
Training Samples: 1,055
Validation Samples: 225
Test Samples: 230
Augmentation: LLM-based data augmentation using Google Gemini

Training Hyperparameters

Learning Rate: 5e-05
Batch Size: 8
Gradient Accumulation Steps: 2
Epochs: 3
Weight Decay: 0.01
Warmup Ratio: 0.1
Optimizer: AdamW
Loss Function: Cross-Entropy with class weights

Training Infrastructure

Framework: HuggingFace Transformers
Hardware: Apple Silicon (MPS) / CPU
Mixed Precision: FP16 (when GPU available)

🎯 Intended Use

Primary Use Cases

Sentiment analysis of Singapore bank reviews
Star rating prediction for financial services
Customer feedback classification
Banking app review analysis

Out-of-Scope Use

Reviews for non-Singapore banks (model trained specifically on Singapore digital banks)
Non-English reviews
General sentiment analysis (domain-specific model)

⚖️ Limitations and Bias

Known Limitations

Trained on Singapore digital bank reviews only
May not generalize well to traditional banks
Performance may vary on very short or very long reviews
English language only

Potential Biases

Data collected from Apple App Store only
May reflect platform-specific sentiment patterns
Class imbalance handled via weighted loss and augmentation

📈 Evaluation Results

The model was evaluated on a held-out test set with the following results:

Accuracy: 41.30%
Average prediction error: 1.02 stars

See confusion matrix and detailed per-class metrics in the training notebook.

🔧 How to Use

For Inference

python

from transformers import AutoModelForSequenceClassification, AutoTokenizer
import torch

# Load model and tokenizer
model = AutoModelForSequenceClassification.from_pretrained("ajiayi/bert-full-finetuned-singapore-digi-banks")
tokenizer = AutoTokenizer.from_pretrained("ajiayi/bert-full-finetuned-singapore-digi-banks")

# Predict
def predict_rating(text):
    inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=128)
    model.eval()
    with torch.no_grad():
        outputs = model(**inputs)
        prediction = outputs.logits.argmax(dim=1).item()
    return prediction + 1  # Convert 0-indexed to 1-5 stars

# Example
review = "Great banking app! Very user-friendly interface."
stars = predict_rating(review)
print(f"Predicted: {stars} stars")

📚 Citation

bibtex

@misc{full-finetuned-singapore-banks,
  author = {Your Name},
  title = {Full Fine-Tuned DistilBERT for Singapore Bank Review Classification},
  year = {2025},
  publisher = {HuggingFace},
  howpublished = {\url{https://huggingface.co/ajiayi/bert-full-finetuned-singapore-digi-banks}}
}

📄 License

This model is released under the Apache 2.0 License.

🙏 Acknowledgments

Base model: DistilBERT by HuggingFace
Training framework: HuggingFace Transformers

Generated: 2025-10-12 00:17:45