duogatech/qwen2.5-7b-deeplearning-hw2-demo

3. Evaluation

• Notebook: task2-eval.ipynb
• Description: Loads the fine-tuned model (from Hugging Face Hub in this case) and the base Qwen2.5 model, then compares their outputs on sample legal questions.
• Execution: Run the cells sequentially. It will download the models from Hugging Face.
• Dependencies: transformers, torch.
• Inputs: Model identifiers from Hugging Face Hub (or local paths if not uploaded).
• Outputs: Printed comparisons of answers from the fine-tuned model and the base model for sample legal questions.