How do I train my own data?

[godisme1220]

Hello, beginner here, thank you for this amazing project, how can I train with my own data?

I checked the code in Colab, and it seems that this code does not only require the input of photos but also the "CCV2017_JTLEE_gtlines_all" file. I checked the data in this folder and found that there are a lot of them. txt file, the .txt file has a parameter that I don't understand, which seems to be the dimensions of the photo.

So here comes the question, how do I get the same information as the .txt file inside "CCV2017_JTLEE_gtlines_all" from my own photos?

And I also encountered a problem, I use the Colab file you provided, it works fine, but inside the "result/reproduce/visualize" folder, I don't see any pictures, they are all named 0,1,2,3. empty folder for ..., am I doing something wrong?