Minimum size for training
Opened this issue · 2 comments
ramonhollands commented
Hi Shihan Liu,
What is the reason there is a minimum input size for the model? In my experiments 192x192.
What would be needed to adapt the model to handle smaller sizes, eg 96x96?
Best regards,
Ramon
LSH9832 commented
see params/train/train_settings.yaml
input_size: [672, 672] # image input size for model
multiscale_range: 5 # real_input_size = input_size + randint(-multiscale_range, multiscale_range) * 32
so if you want to use a very small size, just set multiscale_range to zero
verrassendhollands commented
Thanks for your answer!
I fixed it by preventing random_resize (trainer.py) on small inputs, but this is the improved way.