Model results

Roboflow Model Leaderboard

Choose a shared evaluation protocol to compare model scores. Versions, sources and configurations stay separate.

Atlas snapshot captured 10/19/2026, 10:20:28 AM UTC. This is an import timestamp, not the evaluation date. Coverage may be incomplete.

Official source

← All evaluation groups

Roboflow Model Leaderboard · mAP at IoU 0.75, large objects

Source: Roboflow Model Leaderboard (reviewed publisher results)
Version: COCO 2017 (publisher version unspecified)
Measurement date: 2026-01-15
Metric: fraction · Higher is better
Evaluation scope: COCO 2017 object detection; large objects; publisher-reported mAP at evaluation IoU 0.75; source fractions on 0–1 scale; inference parameters {"num_queries":null,"num_select":null,"resolution":700,"threshold":0}; linked evaluator loads val2017/instances_val2017, but live table does not identify its split or dataset digest; declared model variants, not verified execution checkpoint hashes; executed dependency versions, hardware, precision, run count and uncertainty not supplied; publisher run-record date, not release date; size boundaries and exact aggregation implementation unverified; Unknown release date for the January 15 evaluated checkpoint. Version 1.4.0 on January 22 2026 explicitly introduces new L/XL/2XL checkpoints. Do not silently assign those later checkpoint identities to earlier runs.

Positions are withheld because at least one result lacks explicit independent provenance or a documented evaluation scope. Missing metadata is shown as unreported; matching imported metadata is not independent verification of the protocol.

RF-DETR-XL
roboflow/rf-detr-xl
0.8141669891342712 fraction
View modelEvidence model-card-reported