This project is a Django-based web application that integrates a machine learning model for generating recipes based on a list of ingredients. The model, fine-tuned from the T5-small architecture, is hosted on Hugging Face and utilizes the RecipeNLG dataset for training.
- Recipe Generation: Generate recipe titles, ingredients, and instructions based on a list of ingredients provided by the user.
- Model Training: Fine-tunes the T5-small model on a custom dataset of recipes.
- GPU Acceleration: Utilizes GPU acceleration for efficient model training in Google Colab.
- Google Drive Integration: Load and save model checkpoints directly from Google Drive.
- Fine-Tuned Model: The output from the fine-tuned T5-small model is hosted on Hugging Face.
- Training Dataset: The T5-small model is trained on the RecipeNLG dataset, available at RecipeNLG and Kaggle Dataset. The dataset provides diverse recipe data that enables the model to generate accurate and varied recipe outputs.
Clone the Hugging Face Repository:
Move the Cloned Repository:
- Place the ecochef-recipe-generation folder into the appecochef directory of your Django project. The final structure should be: ecochefmain/appecochef/ecochef-recipe-generation/....
- Python 3.7 or higher
- Django 3.2 or higher
- Libraries:
transformersdatasetspandassentencepieceacceleratetorch
git c
5199
lone https://github.com/yourusername/your-repo-name.git
cd your-repo-name