How to build a deep learning environment running Python in Docker container

Check virtualization in Task Manager, if it is enabled, if it is disabled, restart the computer and enable it in Bios.

Install Docker Desktop and enable the WSL feature.

Open PowerShell

Create an image

docker run --rm -it quay.io/azavea/raster-vision:pytorch-latest /bin/bash
docker images

Create a container based on the image id

docker create -it --name [name] [image id]
docker start [name]

Entering the container

docker exec -it [container id] /bin/bash

Transfer files between docker container and local machine

docker cp [local path] container id:[path inside container]
docker cp D:\docker\raster-vision\myPY\test.py 7837a6060f9d:/opt/src/code/test.py
docker cp container id:[path inside container] [local path] 

View all docker images
View all containers docker ps -a

Expected more than 1 value per channel when training, got input size torch.Size([1, 256, 1, 1])

batch_size=1 results in
Change it to an even number or set the parameter drop_last=True in the torch.utils.data.DataLoader class or in a class you create that inherits from DataLoader to discard data that is less than one batch_size.

urllib.error.URLError: <urlopen error [Errno 111] Connection refused>

Network problem.
Download locally and then upload or use a VPN.

This is the end of this article about how to build a deep learning environment with Python using Docker containers. For more information about how to build a deep learning environment with Python using Docker containers, please search for previous articles on 123WORDPRESS.COM or continue to browse the following related articles. I hope you will support 123WORDPRESS.COM in the future!