Skip to content

Latest commit

 

History

History
52 lines (35 loc) · 3.38 KB

Apple_Detection_Swift-YOLO_192.md

File metadata and controls

52 lines (35 loc) · 3.38 KB

Apple Detection - Swift-YOLO

English | 简体中文 在Colab中打开

版本: 1.0.0

任务 Object Detection

算法: Swift-YOLO

数据集: Apple

类别 apple

Apple Detection

The model is a Swift-YOLO model trained on the apple detection dataset.

网络架构

类型 批次 形状 备注
输入 image 1 [192, 192, 3] The input image should be resized to 192x192 pixels.
输出 bbox 1 [2268, 6] The output is a 2268x6 tensor, where 2268 is the number of candidate boxes and 6 is [x, y, w, h, score, [class]]

基准测试

框架 精度 mAP(%) Flops(M) Params(M) Inference(ms) 下载 作者
PyTorch FLOAT32 91.50 90.564 0.699 - 链接 Seeed Studio
ONNX FLOAT32 89.50 - 0.699 - 链接 Seeed Studio
TFLite FLOAT32 89.50 - - - 链接 Seeed Studio
TFLite INT8 89.30 - - 688.0(1) 链接 Seeed Studio
TFLite(vela) INT8 89.30 - - 45(2) 链接 Seeed Studio

表格注释:

  • Evaluation Parameters: Confidence Threshold: 0.001, IoU Threshold: 0.55, mAP Eval IoU: 0.50..*
  • 框架: 用于推断模型的深度学习框架.
  • 精度: 用于训练模型的数值精度.
  • 指标: 用于评估模型的指标.
  • 推理时间(毫秒): 模型的推理时间(以毫秒为单位).
    • 1: xiao_esp32s3.
    • 2: grove_vision_ai_we2.
  • 链接: 模型的链接.
  • 作者: 模型的作者.

使用指南

许可证

MIT