search

Qengineering / YoloV5-ncnn-Raspberry-Pi-4

YoloV5 for a bare Raspberry Pi 4
https://qengineering.eu/deep-learning-examples-on-raspberry-32-64-os.html
BSD 3-Clause "New" or "Revised" License
52 stars 5 forks source link

ncnn::Layer #1

Closed huyunlei closed 2 years ago

huyunlei commented 3 years ago

Dear author, I can run yolov3, but I can't run yolov5. The error is: ncnn:: layer is not defined. It seems that there is no definition. Can you give me some advice. image

Qengineering commented 3 years ago

Dear @huyunlei,

YoloV5 has an user defined layer, which is not the case in YoloV3 or YoloV4. line 463 in YoloV5.cpp

yolov5.register_custom_layer("YoloV5Focus", YoloV5Focus_layer_creator);

My advise, make absolutely sure you have the latest ncnn version on your machine. (GitHub version 20210322), Make also sure you use the same flags during compilation and linking as in the YoloV5.cbp file, even if you not using code::blocks.

Qengineering commented 3 years ago

Just updated the Code::Blocks project file for the new ncnn version.

huyunlei commented 3 years ago

@Qengineering 作者您好，emm，您可以看一下这个 https://github.com/Tencent/ncnn/issues/1376 用这个也可以解决。 Hello author, emm, you can take a look at this https://github.com/Tencent/ncnn/issues/1376 It can also be solved with this.

huyunlei commented 2 years ago

这是来自QQ邮箱的自动回复邮件。   您好，您发来的邮件已经收到，我会尽快给您回复。