Instruments prediction

[zapaishchykova]

Hi!
First of all, thank you for the great work!
Secondly, I have a question: I have changed the model to detect instruments, like following:
model = get_model(model_path, model_type='UNet11', problem_type='instruments')
As a result in mask i receive a tensor with size 'torch.Size([1, 8, 320, 480])', where i suppose that 8 mean a mask per each instrument. But what i see is that masks 1,3,5,7 are the same as well as 2,4,6,8. ( so it looks like it just make two masks instead of 8) What I'm possibly doing wrong and how to get labels for each instrument?

[tonyreina]

Were you able to figure this out? When I perform inference I'm getting negative and positive numbers for the pixel values. The binary example works, but not sure how to interpret the parts or instruments model outputs. Would anyone have an example of how to post-process those predictions? Thanks.

[WELLBEINGLWB]

Have you figured it out? I encountered the same issue. Thanks in advance for your response.