问题
GCP gives a general overview of what's installed in Deep Learning VMs, but seeing the actual shell scripts would make it easier to determine the exact differences between VM images, debug any deployment issues, and create derivative images.
Someone already asked about the Dockerfiles for Deep Learning Containers, but I figured I'd repeat the question to increase the odds of it getting answered.
回答1:
You can create and set up a local deep learning container. Have a look at the official documentation Getting started with a local deep learning container, but keep in mind that this product is in a pre-release state and might change or have limited support. Here you can find container images to choose from. Unfortunately, there's no source code.
Meanwhile, you can follow instructions from @John Hanley: "You can pull a container image and show the "history" for the container. This shows you how it is built and what the original starting image. This does not mean that you access the original image. You can add to and remove parts of the image. You can also export an image to a tar archive file, modify and then reimport".
来源：https://stackoverflow.com/questions/60027201/where-can-i-find-the-source-code-for-the-google-cloud-platform-deep-learning-vm