Training steps
Malaika68 opened this issue · comments
Hey can you please help me with the training steps? how can I use my dataset and its labels?
Hi Malaika,
you'll need to implement a torch.utils.data.Dataset that returns pairs of image and ground truth annotations.
Then you can change the get_dataset function in data_loading.py to use your custom dataset instead of the pre-configured InriaDataset.
Let me know if that helps!
Cheers,
Konrad
Thank you so much for the help. Can you please tell what type of annotations it should be?
No problem! The annotations should be of the same shape as the imagery (i.e. dense annotations) and contain a 1 for each foreground pixel and a 0 for each background pixel.
In the examples shown in the README, Glacier/Building pixels are the "foreground"
Okay great,,