NVlabs/FasterViT

confuse about use pretrained model to finetune on 1024 x 1024 images

dongshenhan opened this issue · 1 comments

if i change image resolution like (1024,1024),can i also use pretrained model to finetune on my dataset ?

@dongshenhan yes that's the purpose of providing this functionality. Given a strong pre-trained backbone on lower resolution (224x224), we want to enable usage for downstream tasks like detection/segmentation.