ncnn icon indicating copy to clipboard operation
ncnn copied to clipboard

按照yolov8的ncnn推理,在web上部署运行很慢,如何优化

Open xla145 opened this issue 1 year ago • 6 comments

detail | 详细描述 | 詳細な説明

使用了ncnn项目中提供的yolov8推理代码,在web上跑推理很慢 ex.input("in0", in_pad);

std::vector<Object> proposals;

// stride 32
{
    ncnn::Mat out;
    ex.extract("out0", out);

这个转化耗时了800ms,如何进一步优化

xla145 avatar Aug 28 '24 06:08 xla145

你转WASM了?

wzyforgit avatar Aug 29 '24 00:08 wzyforgit

是的,转成wasn’t了

xla145 avatar Aug 29 '24 00:08 xla145

是的,转成wasn’t了

CPU是否有跑满?有编译好的html能否给一个让我们玩一玩?

whyb avatar Aug 29 '24 03:08 whyb

web.zip 拿去吧,但是推理很慢,大佬们优化一下

xla145 avatar Aug 29 '24 04:08 xla145

hi, yolov8 examples are updated with full support for detection, segmentation, classification, pose estimation and obb https://github.com/Tencent/ncnn/tree/master/examples

android demo https://github.com/nihui/ncnn-android-yolov8

detailed instruction (zh) https://zhuanlan.zhihu.com/p/16030630352

nihui avatar Jan 08 '25 08:01 nihui

朋友,我能跟你讨论下不

Jamboo220 avatar Aug 26 '25 09:08 Jamboo220