# Yolo-fastest部屬在PI4上
###### tags: `lab`
> 更新日期:2022/4/12

### 目錄
[toc]

---

訓練:
使用darknet cfg檔請參照
https://github.com/LiaoZihZrong/yolo-fastest/blob/main/yolo-fast.cfg

轉出我們的檔案
git clone https://github.com/LiaoZihZrong/weight-Yolo-fastest-model-h5

# darknet to keras ([tf-14版]轉出h5檔)
`python tools/model_converter/convert.py yolo-fast.cfg yolo-fast.weights yolo-fastest.h5`

# h5 to tflite ([tf2.3]轉出tflite檔)
`python tools/model_converter/custom_tflite_convert.py --keras_model_file yolo-fastest.h5 --output_file yolo-fastest.tflite`

---
(raspberry pi 4)
pip install tensorflow-2.4.0-cp37-none-linux_armv7l.whl
pip install numpy --upgrade