⚡️An Easy-to-use and Fast Deep Learning Model Deployment Toolkit for ☁️Cloud 📱Mobile and 📹Edge. Including Image, Video, Text and Audio 20+ main stream scenarios and 150+ SOTA models with end-to-end optimization, multi-platform and multi-framework support.
https://github.com/facebookresearch/segment-anything![微信截图_20230410112431](https://user-images.githubusercontent.com/6490927/230864550-7d241b7b-621c-4ab1-ab19-9a0b1a7a8779.png)
一切都能分割
分割任何模型 (SAM) 根据输入提示(如点或框)生成高质量的对象蒙版,并可用于为图像中的所有对象生成蒙版。它已经在 11 万张图像和 1 亿个掩码的数据集上进行了训练,并且在各种分割任务上具有强大的零镜头性能。
这个已经转换为onnx 怎么用?