Just a try about deploying a classification model on android platform
- deploying a model on android: https://zhuanlan.zhihu.com/p/32342366
- get train and test imgs from Baidu:https://github.com/kong36088/BaiduImageSpider
- install Android Studio: https://blog.csdn.net/u013038616/article/details/75315283
- Get train and test imgs
- Start finetune
- Using onnx to convert pytorch model into caffe2 model
- clone AI Camera Demo from github: git clone https://github.com/bwasti/AICamera.git
- change path of model and labelname
- pytorch, caffe2 and onnx
- python2 and python3 environment
- cuda8.0, ncll, cudnn7.1
- 狗 2. 手机 3. 猫 4. 杯子 5. 花 6. 书桌 7. 人 8. 电视机 9. 电脑 10. 钱包
- 笔记本 12. 眼镜 13. 鼠标 14. 汽车 15. 自行车 16. 火车 17. 椅子 18. 餐桌 19. 沙发 20. 鸟