Skip to content

YOLOv10:

Real-time object detection without NMS

Visit official site
Category AI models

What YOLOv10 is

The model is described as a next-generation object detector that keeps real-time performance while reducing compute redundancy through optimised post-processing and architecture.

What you can do with it

  • Run detection without non-maximum suppression
  • Use pretrained models or trained checkpoints
  • Work with several input image resolutions

Application scenarios

  • A surveillance system detects unusual behaviour in real time
  • A driving system recognises pedestrians and vehicles
  • A production line checks product quality automatically

Pros & cons

✓ What we like

  • Several input resolutions are supported
  • Multiple scales are evaluated on the COCO dataset

! What to watch out for

  • Detection follows the chosen model scale
  • Setup needs a Python environment and dependencies

FAQ

What is removed from post-processing?

The feature list names consistent dual assignment without non-maximum suppression.

Which frameworks are named?

The card lists support for frameworks such as PyTorch.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review