An end-to-end NLP-based emotion classification system built using PyTorch and Streamlit.
The model predicts emotions from user-input text using a BiLSTM neural network trained on the dair-ai/emotion dataset.
- Text preprocessing and custom vocabulary
- BiLSTM-based emotion classifier
- Handles mixed-emotion ambiguity using probability scores
- Interactive Streamlit web interface
- Model persistence for fast inference
- Joy
- Sadness
- Anger
- Fear
- Love
- Surprise
- Python
- PyTorch
- NLP
- BiLSTM
- Streamlit
- Hugging Face Datasets
pip install -r requirements.txt
streamlit run emotion.py