zhangwenguanghikvision / yolov5_jde_tensorrt_for_track Goto Github PK
View Code? Open in Web Editor NEWA multi object tracking Library Based on tensorrt
License: MIT License
A multi object tracking Library Based on tensorrt
License: MIT License
There is a "while" section in the function "void solve(Matrix &m)",
In some situations, the function "int step2()" never return 0, so the "while" is caught in dead loop.
Thanks!
您好!
这里想友好探讨一下:
我这边跑了一下您的项目,检测效果很好。
但是当遇到遮挡或者两个人擦身而过,就容易跟丢,虽然使用了CStracker,但是跟踪效果没有达到fairmot的跟踪效果。
不知道您这边的跟踪效果怎么样?
Hello, do you have the unaccelerated version of yolov5_JDE Python? Could you please give me your contact information for some details about JDE.
I'd like to know about the difference between your jde model and the official model in which the 3 * (5 + classes_num) + embedding dimensions feature directly from the last same convolution simultaneously.
hello,
your project is helpful! thanks a lot!
however, I use TensorRT >= 7.1 with CUDA 10.2, and your project could not run, because the version not match.
could you provide a docker image for your project, it will help a lot!
首先,感谢您共享了代码。
关于CSTrack算法,我比较感兴趣转码后的处理速度如何。
这个算法,我最初也打算用TRT加速推理。只是在CCN模块遇到硬伤AdaptiveAveragePooling2D无法转成onnx。因此,我将这个模型转成JIT script,并用libtorch去推理,包括后处理部分卡尔曼滤波等这些都是直接用libtorch去做矩阵计算。
遗憾的是算法输入1333*800比较大, libtorch gpu推理速度 跟 pytorch 差不多,并没有明显提速。整个算法在数据集MOT17-03 测试集 下大约是5fps(i7 7700+1060显卡)。
我查阅了 一下 libtorch 推理速度不快的原因,貌似也有大佬遇到同样的问题。
https://zhuanlan.zhihu.com/p/363319763
因此,我想大致了解一下去除SAAN 和CCN后,用TRT处理速度是多少?对于这个算法的加速推理和计算还有更好的建议吗?谢谢
如题,百度云下载的wts文件转化为yolov5l_int8.engine模型后,程序能顺利跑起来,但是识别不到人,请问是模型有问题吗?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.