ahczhg/nvidia-dler-r1-7b-demo

NVIDIA DLER-R1-7B Research Demo

This repository contains a Jupyter notebook demonstrating the usage of NVIDIA's DLER-R1-7B-Research model for question answering and code generation tasks.

Overview

DLER-R1-7B is a 7-billion parameter language model developed by NVIDIA for research purposes. This notebook showcases:

• Setting up the model with the Transformers library
• Question answering capabilities (e.g., explaining SDN)
• Code generation abilities (e.g., generating Python KNN implementation)

Features

• Question Answering: Ask technical questions and receive detailed explanations with reasoning traces
• Code Generation: Generate functional Python code with explanations
• GPU Acceleration: Automatically uses CUDA if available for faster inference

Requirements

• Python 3.8+
• CUDA-capable GPU (optional but recommended)
• Dependencies listed below

Installation

1. Clone this repository:

git clone <your-repo-url>
cd <repo-name>

2. Install required packages:

pip install transformers==4.51.3 torch

Usage

Open the Jupyter notebook:

jupyter notebook Untitled0.ipynb

The notebook includes examples of:

1. Technical Question Answering:

   • Example: "What is SDN?"
   • The model provides a detailed explanation with reasoning

2. Code Generation:

   • Example: "Write a Python code for KNN"
   • Generates complete, functional code with explanations

Model Information

• Model: nvidia/DLER-R1-7B-Research
• Size: 7B parameters
• Type: Causal Language Model
• Framework: Hugging Face Transformers

Examples

Question Answering

messages = [
    {"role": "user", "content": "what is SDN?"},
]

Code Generation

messages = [
    {"role": "user", "content": "write a python code for KNN"},
]

Performance Notes

• The model uses Sliding Window Attention
• Requires approximately 15GB of disk space for model files
• Inference time depends on GPU availability and prompt complexity
• Maximum generation tokens set to 10,000 for detailed responses

License

MIT License - see LICENSE file for details

Acknowledgments

• NVIDIA for developing and releasing the DLER-R1-7B-Research model
• Hugging Face for the Transformers library

Citation

If you use this code or the NVIDIA DLER-R1-7B model in your research, please cite:

@misc{nvidia-dler-r1-7b,
  title={DLER-R1-7B-Research},
  author={NVIDIA},
  year={2024},
  publisher={Hugging Face},
  howpublished={\url{https://huggingface.co/nvidia/DLER-R1-7B-Research}}
}

Contributing

Contributions, issues, and feature requests are welcome! Feel free to check the issues page.

Contact

For questions or feedback, please open an issue in this repository.