Fine-tuning Gemma on Climate Data

This project fine-tunes the Gemma language model using a climate question-answer dataset to enhance its domain-specific knowledge in environmental topics. By leveraging Low Rank Adaptation (LoRA), the model is optimized to provide more accurate and relevant responses to questions related to environmental conflicts and climate issues.

What the Project Does

• Loads and Prepares Data: Utilizes a climate question-answer dataset from Kaggle.
• Model Fine-Tuning: Applies LoRA for efficient fine-tuning, enabling the model to learn specific climate-related knowledge.
• Inference Improvement: Enhances the model's ability to generate accurate answers to climate-related questions.
• Model Upload: Provides functionality to save and upload the fine-tuned model to Kaggle for further use and distribution.