streamlit run app.py
docker build -f Dockerfile -t app:latest .
docker run -p 8501:8501 app:latest
docker run --gpus all -itd [docker_image] /bin/bash
docker exec -it [container_id] bash #by ID
cd mnt/c/
docker run --gpus all -itd -v $(pwd):/workspace
mklink /d ~\data_symlinks\linkage $(pwd)
ln -s /workspace ./hico_exp
docker cp container_id:/foo.txt foo.txt
CUDA toolkit is for compiling CUDA application on WSL 2
WSL 2 GPU Paravirtualization no need for NVIDIA Container Toolkit
https://maelfabien.github.io/project/Streamlit/#dockerfile
https://docs.nvidia.com/cuda/wsl-user-guide/index.html#ch02-sub03-installing-wsl2
https://blog.roboflow.com/use-the-gpu-in-docker/