I'm trying to set up the environment to train HF-Net and want to make sure I'm putting things together correctly. It seems the way to create the items in the bdd/dawn_images_vga and bdd/night_images_vga folders is to download the BDD100k images and labels then write a script to parse the labels and copy the dawn and night images into the appropriate folders. Is that right? Is there already a script for this? thanks
Hi and sorry for the delay. We indeed set it up this way, but the relative paths can be changed from the configuration files. Sadly I do not have such script anymore, but it should be fairly straightforward.
I'm trying to set up the environment to train HF-Net and want to make sure I'm putting things together correctly. It seems the way to create the items in the bdd/dawn_images_vga and bdd/night_images_vga folders is to download the BDD100k images and labels then write a script to parse the labels and copy the dawn and night images into the appropriate folders. Is that right? Is there already a script for this? thanks