Configuration and Dependencies

Configuration and Dependencies

• requirements.txt: PyTorch >=2.6, tqdm >=4.66, Transformers ==5.22.2, safetensors >=0.4.
• requirements-notebook.txt: above plus Notebook, ipykernel and nbformat.
• Python 3.11+ is required by the pinned Transformers 5.22.2 release; Python 3.12 recommended.
• CUDA-enabled PyTorch is installed separately from the official platform-specific wheel selector.
• Model defaults to Qwen/Qwen2.5-0.5B-Instruct; download/cached weights can be large and may require license / access review for alternative models.
• The program does not load .env files or call a cloud inference API; model weights are loaded locally in the worker process.

Validate compatibility after installing:

python -m pip check
python -c "import torch; print('CUDA available',torch.cuda.is_available(),'count',torch.cuda.device_count())"

See Getting Started for execution examples and Troubleshooting for failure cases.