Recently, I just used Docker Desktop to run deep reinforcement learning code with an image. Then, there was a particularly strange bug. I used docker load to import the tar package provided by my senior, and after importing the image, I found that the image size had increased by 20G. Others’ imported image size was only 20G, but mine showed 40G. Then, I checked the image size through docker system df, and it was only a little over 20G, which was normal. However, docker images still showed 40G, which was very strange.
Today, I downloaded the ray-ml image provided by RAY from Docker Desktop and found that the downloaded image was also 20G larger. The original official size should be around 10G, but the one I downloaded was 30G.
Has anyone encountered a similar problem? Do you know how to solve it? Thank you very much.