Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/Baiyuetribe/ncnn-models
awesome AI models with NCNN, and how they were converted ✨✨✨
https://github.com/Baiyuetribe/ncnn-models
deep-learning gfpgan image-classification nanodet ncnn ncnn-model pytorch rife yolov5 yolov7 yolox
Last synced: 2 months ago
JSON representation
awesome AI models with NCNN, and how they were converted ✨✨✨
- Host: GitHub
- URL: https://github.com/Baiyuetribe/ncnn-models
- Owner: Baiyuetribe
- License: mit
- Created: 2022-05-20T02:28:29.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2023-07-07T15:32:23.000Z (over 1 year ago)
- Last Synced: 2024-11-07T17:33:51.672Z (2 months ago)
- Topics: deep-learning, gfpgan, image-classification, nanodet, ncnn, ncnn-model, pytorch, rife, yolov5, yolov7, yolox
- Language: C++
- Homepage: https://github.com/Baiyuetribe/ncnn-models
- Size: 150 MB
- Stars: 253
- Watchers: 10
- Forks: 29
- Open Issues: 15
-
Metadata Files:
- Readme: README-CN.md
- License: LICENSE
Awesome Lists containing this project
- awesome-artificial-intelligence - NCNN Models - awesome AI models with NCNN, and how they were converted sparklessparklessparkles. (Other awesome AI lists)
- awesome-yolo-object-detection - Baiyuetribe/ncnn-models - models?style=social"/> : awesome AI models with NCNN, and how they were converted ✨✨✨ (Lighter and Deployment Frameworks)
- awesome-yolo-object-detection - Baiyuetribe/ncnn-models - models?style=social"/> : awesome AI models with NCNN, and how they were converted ✨✨✨ (Lighter and Deployment Frameworks)
- awesome - ncnn-models - awesome AI models with NCNN, and how they were converted. (OOM-Leaks-Crash / AR-VR)
- awesome - ncnn-models - awesome AI models with NCNN, and how they were converted. (OOM-Leaks-Crash / AR-VR)
README
##
ncnn models
The collection of pre-trained AI models, and how they were converted, deployed.
![](docs/images/logo.png)
### 关于
ncnn框架借助vulkan实现了全平台部署,我们通过pytorch、tensorflow、飞桨等预训练模型,然后转换成ncnn通用模型,实现Windows、mac、linux、安卓、ios、WebAssembly 以及uni-app的最终部署。然而模型转换并不是一键的,需要手动处理。为了拓展ncnn的边界应用,我们建立此仓库,欢迎提交接收任何转换成功或失败的案例。
### 如何参与贡献?
fork代码,然后按照如下格式提交,最好是c++20 的最小demo,不要内嵌并发。
yolov5 # 项目名称
- models/xx.bin or xx.param # model
- input.png # 输入
- out.png # 输出
- README.md # 模型推理介绍、转换步骤、相关案例
- convert.py # 从pytorch等模型转换具体复现代码灵感来自[ailia](https://github.com/axinc-ai/ailia-models),鉴于接收失败案例为主,因此失败案例优先排序,成功案例分类排序。
### 一些代表
- [nihui](https://github.com/nihui) ncnn作者
- [飞哥](https://github.com/feigechuanshu) ncnn安卓系列
- [EdVince](https://github.com/EdVince) ncnn自然语言系列
- [baiyue](https://github.com/Baiyuetribe/paper2gui) ncnn在PC桌面GUI系列
- [670***@qq.com](https://ext.dcloud.net.cn/plugin?id=5243) ncnn在uni-app中的应用
- [nihui](https://github.com/nihui/ncnn_on_esp32) ncnn在嵌入式设备上的应用
- [nihui](https://github.com/nihui/ncnn-webassembly-yolov5) ncnn在wasm实现案例### 跨设备的意义
> 维护一次代码,多设备可用。
- 借助vulkan,无需繁重的cuda驱动
- 借助wasm实现任意设备的模型部署
- 借助uni-app可绕过java实现多端部署