Unverified75% confidenceFactTime unknown

YOLOv8-nano模型在中端NPU上可实现30FPS以上的推理帧率

1

Sources

75%

Confidence

Long-term

Relevance

6/1/2026

First Seen

Sources

戴森Find+Follow空气净化风扇评测：AI摄像头追踪人员位置精准送风

rss

Related Entities

YOLOv8-nanomodel NPUconcept

Related Claims

Unverified轻量化模型通常采用YOLOv8-nano、MobileNet等经剪枝量化蒸馏处理的网络，在低分辨率下实现每秒30帧以上实时推理77% similar UnverifiedYOLOv8 模型分为 nano 到 x 多个规格，YOLOv8s 在边缘设备上可实现实时推理77% similar UnverifiedMoS-TTS-Nano可能采用了非自回归架构来平衡质量与推理速度65% similar UnverifiedYOLOv8 模型家族从 YOLOv8n（约 3.2M 参数）到 YOLOv8x（约 68M 参数）覆盖从树莓派到高性能 GPU 的硬件谱系64% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/25905

API

curl https://kongchang.com/api/v1/knowledge/claims/25905

MCP

get_claim(id=25905)