A LLaMA 2 model is implemented in the PyTorch framework from scratch.
- Ensure the required libraries are installed on your system.
pip install -r requirements.txt
- Download model weights with the official llama repository or from LLaMA 2 7B on Hugging Face. Make sure to include the following:
pytorch-llama
├── llama-2-7b
│ ├── checklist.chk
│ ├── consolidated.00.pth
│ └── params.json
├── tokenizer.model