multi-class datasets issues
wangqingl opened this issue · 2 comments
Hello, My dataset has ten categories of changes, and I have defined labels as numbers from 0 to 9. The modifications I made in the code are shown in the following figure. Is this feasible?
When testing the effectiveness of the trained model, I made modifications to the tested code, as shown in the following figure. I deleted the "* 255" in the code. But the test result is a pure black single channel image, which means that the trained model has no effect. How can this be solved? Looking forward to your reply.
@wangqingl If you have 10 classes you need to write separate visualizer to visualize all the classes. Since, you removed pred * 255 your values the output.png will be between 0 and 9 - hence you see dark (or black) output. You need to write the visualizer such that it will assign different color values for each predicted output class.
@wangqingl Did you able to manage this issue?