Intellihack_TeamX_Task3

Submission for IntelliHack 5.0 LLM Task 1 by TeamX.

Contents

• IntelliHack_LLM_Submission.ipynb: Notebook with training, evaluation, inference, and report.
• Model Files:
  • Model files : Google Drive Link to qwen_finetuned

Instructions

1. Open IntelliHack_LLM_Submission.ipynb in Google Colab.
2. Mount Drive: from google.colab import drive; drive.mount('/content/drive').
3. Install dependencies: !pip install torch==2.6.0+cu124 -f https://download.pytorch.org/whl/torch_stable.html; !pip install unsloth==2025.3.9 transformers==4.48.3 datasets==2.19.0.
4. Run cells to load model and test inference.
5. Model files are in Drive (links above) or upload locally by downloading them.

Notes

• Trained on 5 DeepSeek .md files, 43 train/11 test examples.
• Uses Qwen2-0.5B with LoRA, 4-bit quantization for efficiency.