Vision Analysis
Back to Leaderboard

DEIMv2

deimv2In LibreYOLO

Intellindust AI Lab · arXiv 2025

Variants8
Parameters0.5M - 50.3M
Best Paper mAP57.8%
LicenseApache-2.0

Model Variants

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

ModelParamsGFLOPsPaper mAPBest FPS
DEIMv2-Atto0.5M0.823.8%
138 FPS
NVIDIA RTX 5070 Ti · TensorRT FP32
Details
DEIMv2-Femto1.0M1.731%
119 FPS
NVIDIA RTX 5070 Ti · TensorRT FP32
Details
DEIMv2-Pico1.5M5.238.5%
80 FPS
NVIDIA RTX 5070 Ti · ONNX Runtime FP32
Details
DEIMv2-N3.6M6.943%
78 FPS
NVIDIA RTX 5070 Ti · TensorRT FP16
Details
DEIMv2-S9.7M25.650.9%
53 FPS
NVIDIA RTX 5070 Ti · TensorRT FP32
Details
DEIMv2-M18.1M52.253%
50 FPS
NVIDIA RTX 5070 Ti · TensorRT FP32
Details
DEIMv2-L32.2M96.356%
43 FPS
NVIDIA RTX 5070 Ti · TensorRT FP32
Details
DEIMv2-X50.3M151.657.8%
38 FPS
NVIDIA RTX 5070 Ti · TensorRT FP32
Details

Architecture

Type

transformer

Backbone

DINOv3-distilled ViT

Neck

HybridEncoder

Head

DEIM

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