Machine learning experiment.
This project uses the NVIDIA Docker image, which needs the nvidia-container-toolkit
installed on the host. On Ubuntu:
sudo apt install nvidia-container-toolkit
To run, use
docker build . -t confession
docker run --rm --gpus all -it -p 5000:8888 confession bash
Check that CUDA is working by calling python3
and executing
import torch
torch.cuda.is_available() # True
The -p 5000:8888
argument on the docker run
command allows us to do
jupyter notebook --ip 0.0.0.0 --port 8888 --allow-root &
in the container
and go to the http://127.0.0.1:5000
URL in a host browser to get Jupyter.