PyTorch Lightning + Hydra. A very general, feature-rich template for rapid and scalable ML experimentation with best practices. 🔥
Перейти к файлу
hobogalaxy 8c77b7f3eb update dockerfile simple 2021-04-08 21:03:22 +02:00
custom update 2021-04-08 16:08:57 +02:00
simple update dockerfile simple 2021-04-08 21:03:22 +02:00
.gitignore Dev (#118) 2021-04-05 19:39:56 +02:00
README.md update dockerfiles 2021-04-08 14:38:13 +02:00

README.md

Dockerfiles

  • simple - build from official nvidia pytorch image, contains: apex, miniconda
  • custom - build from official nvidia cuda image, contains: apex, initialized miniconda environment, ready to run template, allows for specifying dependiecies (versions of cuda, python, pytorch)

Make sure you have installed the NVIDIA driver >= 361.93 and Docker >= 19.03
https://github.com/NVIDIA/nvidia-docker/wiki/Installation-(Native-GPU-Support)#prerequisites

cd simple
docker build -t lightning-hydra .
docker run --gpus all -it --rm -v /home/USER/Desktop:/workspace/Desktop lightning-hydra