search

dog-qiuqiu / Yolo-Fastest

:zap: Based on yolo's ultra-lightweight universal target detection algorithm, the calculation amount is only 250mflops, the ncnn model size is only 666kb, the Raspberry Pi 3b can run up to 15fps+, and the mobile terminal can run up to 178fps+
Other
1.98k stars 427 forks source link

Android deployment #22

Open zxj11838 opened 3 years ago

zxj11838 commented 3 years ago

Can I use this project（Android_NCNN_yolov4-tiny） to deploy？

wwdok commented 3 years ago

I am curious about this too, do you mean just replace with yolo-fastest bin and param ?

zxj11838 commented 3 years ago

@wwdok If it’s a model you trained yourself, You need to modify these three places to your corresponding information. 1、Box.java private static String[] labels
2、jni_interface.cpp YoloV5::detector = new YoloV5(mgr,"yolo-fastest.param","yolo-fastest.bin"); 3、YoloV5.h int input_size

jiangjiaxiaotianxia commented 3 years ago

I modify these three places,and successfully reduce the app in my phone .But when detecting ,there is no any box in the picture,Why?

jiangjiaxiaotianxia commented 3 years ago

@zxj11838