Open DucTriCE opened 1 day ago
and one more question, if i'm using co_dino_vit/co_dino_5scale_vit_large_coco.py for object detection training task, according to the link for downloading pretrained to put it 'load_from', which model is correct in the attached image below
where can i find the "models/swin_large_patch4_window12_384_22k.pth", I'm using the base 'co_dino_5scale_swin_large_3x_coco.py' to train object detection for custom dataset, in my own config, I already have the load_from = 'checkpoints/co_dino_5scale_swin_large_3x_coco.pth'. However when running it requires the abovementioned one. Can you help me ?