Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB
💎1MB lightweight face detection model (1MB轻量级人脸检测模型)
https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB
arm face-detection inference mnn ncnn
Last synced: about 2 months ago
JSON representation
💎1MB lightweight face detection model (1MB轻量级人脸检测模型)
- Host: GitHub
- URL: https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB
- Owner: Linzaer
- License: mit
- Created: 2019-10-10T07:40:48.000Z (about 5 years ago)
- Default Branch: master
- Last Pushed: 2023-12-29T15:39:12.000Z (12 months ago)
- Last Synced: 2024-10-15T00:21:08.870Z (2 months ago)
- Topics: arm, face-detection, inference, mnn, ncnn
- Language: Python
- Homepage:
- Size: 36.4 MB
- Stars: 7,159
- Watchers: 191
- Forks: 1,538
- Open Issues: 128
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-lists - Ultra-Light-Fast-Generic-Face-Detector
- awesome-ncnn - Ultra-Light-Fast-Generic-Face-Detector-1MB
- awesome-list - Ultra-Light-Fast-Generic-Face-Detector-1MB - 1MB lightweight face detection model. (Computer Vision / Classification & Detection & Tracking)
- StarryDivineSky - Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB - 快速-通用-人脸检测器-1MB，超轻量级人脸检测模型。模型大小方面，默认FP32精度（.pth）文件大小为1.04~1.1MB，推理框架int8量化大小约为300KB。就模型的计算量而言，320x240的输入分辨率约为90~109 MFlops。该模型有两个版本，version-slim（网络骨干简化，稍快）和version-RFB（带有修改后的RFB模块，精度更高）。提供320x240和640x480不同输入分辨率的Widerface训练预训练模型，以更好地在不同应用场景下工作。支持 onnx 导出，便于迁移和推理。提供 NCNN C++ 推理代码。提供MNN C++推理代码，MNN Python推理代码，FP32/INT8量化模型。提供 Caffe 型号和 onnx2caffe 转换代码。Caffe python 推理代码和 OpencvDNN 推理代码。 (人像\姿势\3D人脸 / 网络服务_其他)