This project predicts a user's MBTI personality type based on a text input and matches it with an F1 driver who shares the same personality type.
It uses a BERT-based model for accurate natural language understanding and is wrapped in an interactive Dash web app UI for a smooth and stylish experience.
Dataset : Link
Colab Notebook : Used for training the model using GPU. Link
Model: Fine-tuned using BERT (bert-base-uncased) to classify MBTI types from text inputs.
Dataset: A large corpus of personality-labeled text was used for training Link.
Training: The model was trained on Google Colab for convenience and GPU acceleration.
Frontend: Built using Plotly Dash, allowing real-time predictions and a visual UI with images of real F1 drivers.
Model weights : model.safetensors file is very large for GITHUB. You can train your own model to acquire the file, or use the deployed link.
Accepts a sentence or paragraph written by a user.
Processes it through a fine-tuned BERT model.
Predicts one of 16 MBTI personality types.
Matches that personality to a real-life F1 driver.
Displays the prediction and the corresponding driver's name + image.
