RF-DETR Segmentation. Benchmarks, Inference, Training | Live Coding + Q&A (Jan 29th)

In this live stream I introduce RF-DETR Segmentation and explain how to use the model for real-time instance segmentation in practice. I review benchmark results on MS COCO and compare accuracy and latency with other real-time segmentation models, including recent YOLO26 variants. The focus stays on practical performance tradeoffs across model sizes rather than on release details. I then show how to run RF-DETR-Seg inference in Python and through Roboflow Inference using production-ready workflows. I demonstrate how to train and fine-tune RF-DETR-Seg on a custom instance segmentation dataset starting from pretrained checkpoints. This stream targets engineers working on real-time computer vision systems for segmentation and object detection. Resources: - 📓 RF-DETR instance segmentation model training notebook: https://colab.research.google.com/github/roboflow-ai/notebooks/blob/main/notebooks/how-to-finetune-rf-detr-on-segmentation-dataset.ipynb - 📓 RF-DETR instance segmentation blogpost: https://blog.roboflow.com/rf-detr-segmentation Stay updated with the projects I'm working on at https://github.com/roboflow and https://github.com/SkalskiP! ⭐