This folder contains several notebooks illustrating how to use DETR for inference as well as fine-tuning on custom data.
- inference with DetrForObjectDetection on an image
- fine-tuning DetrForObjectDetection on a custom dataset
- evaluating DetrForObjectDetection on COCO
- inference with DetrForSegmentation on an image
- fine-tuning DetrForSegmentation on a custom dataset
All DETR checkpoints can be found on the hub.
Update: we now also support Deformable DETR, which improves upon the original DETR in terms of convergence.