ppogg / yolo-streaming Goto Github PK

View Code? Open in Web Editor NEW

73.0 0.0 49.0 65.91 MB

Push-pull streaming and Web display of YOLO series

Python 99.02% HTML 0.98%

yolov3 yolov4 yolov5 yolox stream web dnn onnxruntime flask

yolo-streaming's Introduction

Hi there! I'm Xiangrong Chen.

Hi, I’m Xiangrong Chen, a student who likes to record life, and here is my Garden of Eden~

🔭 When I was in undergraduate, I teamed up with my classmates to attend competition. At that time I was exposed to machine learning and was fascinated by it.🤖
🌱 During graduate school, I chose deep learning as my research direction. I believe that there will be many interesting people and things on this way.🤣
👯 During the second year of graduate school, I have practiced in Guangzhou and recorded my own interesting experiments during the period.
💡 I like to program by python and C++
🤣 Hope to meet more friends. Nice to meet you~ 😁

Things I code with

Open source projects

🎁 Projects	⭐ Stars	📚 Forks	🛎 Issues	📬 Pull requests
YOLOv5-Lite
YOLO-Streaming
Retinaface_Ghost
ncnn-yolov4-int8

I like listening to music, playing ball, and like to adventurous more. 🎼🎸🎶 → Spotify Playing 🎧👀

At the same time, I also have other blogs, welcome to pay attention!💻📌

➡️ Xiangrong Chen's home page in zhihu💛
➡️ Xiangrong Chen's home page in CSDN💛

Connect with me and Grab a Coffee! 🙋‍♀️🥤

This README file is generated every 3 hours!
Last refresh: Monday, 14:07 CEST
Create your own here!

yolo-streaming's People

Contributors

Stargazers

Forkers

matu-46177 jiangyuewu darkknight2223 chaucerg zlszhonglongshen cccczl give-me-yasoo huanggua123456 shizhanhao luoyizhi516 okstill flowold trisword163 zbh10

yolo-streaming's Issues

是否可以支持多路？

你好，请问一下该代码是否支持多路实时显示

Error reported about `outs[i].reshape`

关于outs[i].reshape的报错

作者您好，
我从github clone下你的代码，所有环境均已配置完成，但是运行单个推理文件python v5_dnn.py的时候会报错：

Traceback (most recent call last):
  File "D:\Lecter\Downloads\ml-proj\YOLO-Streaming\v5_dnn.py", line 118, in <module>
    yolonet.v5_inference(frame)
  File "D:\Lecter\Downloads\ml-proj\YOLO-Streaming\v5_dnn.py", line 38, in v5_inference
    outs = self.detect(frame)
  File "D:\Lecter\Downloads\ml-proj\YOLO-Streaming\v5_dnn.py", line 98, in detect
    outs[i] = outs[i].reshape(bs, self.na, self.no, ny, nx).transpose(0, 1, 3, 4, 2)
TypeError: 'tuple' object does not support item assignment

指向的是v5_dnn.py文件中：

        for i in range(self.nl):
            bs, _, ny, nx = outs[i].shape  # x(bs,255,20,20) to x(bs,3,20,20,85)
            # outs[i] = outs[i].view(bs, self.na, self.no, ny, nx).permute(0, 1, 3, 4, 2).contiguous()
            outs[i] = outs[i].reshape(bs, self.na, self.no, ny, nx).transpose(0, 1, 3, 4, 2)
            if self.grid[i].shape[2:4] != outs[i].shape[2:4]:
                self.grid[i] = self._make_grid(nx, ny)

            y = 1 / (1 + np.exp(-outs[i]))  ### sigmoid
            ###其实只需要对x,y,w,h做sigmoid变换的， 不过全做sigmoid变换对结果影响不大，因为sigmoid是单调递增函数，那么就不影响类别置信度的排序关系，因此不影响后面的NMS
            ###不过设断点查看类别置信度，都是负数，看来有必要做sigmoid变换把概率值强行拉回到0到1的区间内
            y[..., 0:2] = (y[..., 0:2] * 2. - 0.5 + self.grid[i]) * int(self.stride[i])
            y[..., 2:4] = (y[..., 2:4] * 2) ** 2 * self.anchor_grid[i]  # wh
            z.append(y.reshape(bs, -1, self.no))

中的outs[i] = outs[i].reshape(bs, self.na, self.no, ny, nx).transpose(0, 1, 3, 4, 2)

报错类型为TypeError: 'tuple' object does not support item assignment

麻烦可以帮我看一下吗

运行无结果

你好，想问一下我下载代码后，想实现在线视频检测，在终端运行 python app.py --model v3_fastest后opencv报错，显示没有接收到视频流，且网页界面如下图所示，我也没有找到在哪里输入视频流，请问一下我是哪里做错了吗？