RT-DETRv4

rtdetrv4In LibreYOLO

Intellindust AI Lab · arXiv 2025

Paper Code

Variants4

Parameters10.0M - 62.0M

Best Paper mAP57%

LicenseApache-2.0

Model Variants

All variants benchmarked on COCO val2017. Click any variant for the full hardware breakdown.

Model	Params	GFLOPs	Paper mAP	Best FPS
RT-DETRv4-S	10.0M	25.0	49.7%	97 FPS NVIDIA RTX 5070 Ti · TensorRT FP32	Details
RT-DETRv4-M	19.0M	57.0	53.5%	87 FPS NVIDIA RTX 5070 Ti · TensorRT FP16	Details
RT-DETRv4-L	31.0M	91.0	55.4%	78 FPS NVIDIA RTX 5070 Ti · TensorRT FP16	Details
RT-DETRv4-X	62.0M	202.0	57%	71 FPS NVIDIA RTX 5070 Ti · TensorRT FP16	Details

Architecture

Type

transformer

Backbone

HGNetv2

Neck

HybridEncoder

Head

DETR

detrnms-freeTrained on COCO train2017

Run any model with one line

LibreYOLO has the best catalogue of state-of-the-art detectors, all behind one MIT-licensed Python API.

from libreyolo import LibreYOLO, SAMPLE_IMAGE # LibreYOLO has the best catalogue of state-of-the-art models. model = LibreYOLO("LibreRFDETRl.pt") # RF-DETR-L (transformer flagship) results = model(SAMPLE_IMAGE, save=True) # run inference, save the annotated image # Swap in any other model, same one-line API (weights auto-download): # LibreYOLO("LibreYOLO9c.pt") # YOLO9-C # LibreYOLO("LibreYOLOXx.pt") # YOLOX-X # LibreYOLO("LibreDFINEx.pt") # D-FINE-X # LibreYOLO("LibreRTDETRr50.pt") # RT-DETR-R50