Hi everyone!
I’m trying to train my own dataset on the HybridNets architecture and I tried looking at the links provided for the dataset to find out the structure of the data but the links were broken.
I looked at some other segmentation models and they seemed to be using two image files each in train_images and train_masks and for object detection, it’s a file with a set of points that make up a polygon. Based on my understanding, for object detection, there are multiple formats to display this encoding.
So now my question is, how should I set up my dataset to train for HybridNets on it?
And how did you find out the answer to this? (so that I can figure it out myself next time)