Docker Hub
hub.docker.com › r › bitnamiRunning your PyTorch app The default work directory for the PyTorch image is /app. You can mount a folder from your host here that includes your PyTorch script, and run it normally using the python command. $ docker run -it --name pytorch -v /path/to/app:/app bitnami/pytorch \ python script.py Running a PyTorch app with package dependencies
PyTorch GPU inference with Docker and Flask :: Päpper's ...
www.paepper.com › blog › postsMay 04, 2019 · In this article, I will show you how to use Docker to serve your PyTorch model for GPU inference and also provide it as a REST API. You can then use AWS API Gateway or any other service to handle authentication and security measures to serve this in production. For the time being, I will simply show you the basics for the PyTorch inference part.
docker image info
docker.pytorch.org › docker_hubDocker images on docker hub; repo tag size last_updated_at last_updated_by; pytorch/conda-cuda: latest: 8178639006: 2020-03-09T20:07:30.313186Z: seemethere: pytorch/conda-cuda-cxx11-ubuntu1604