各模型的py文件中仅包含此模型,main.py为主文件,compare.py用于比较各个模型(已选取最优参数),plot.py中整合了画图,方便调用
参数--model 选定模型
参数--lr 学习率 可以多填
参数--dropout 可以多填
参数--plot 是否画图
参数--epoch 轮次
示例:
python .\main.py --model googlenet --lr 0.001 0.0005 0.0001 --dropout 0 0.2 0.4 --plot true --epoch 5
内置了五个经典模型以及合适的学习率,运行以下代码即可
python .\compare.py
[1] Y. Lecun, L. Bottou, Y. Bengio and P. Haffner, "Gradient-based learning applied to document recognition," in Proceedings of the IEEE, vol. 86, no. 11, pp. 2278- 2324, Nov. 1998, doi: 10.1109/5.726791.
[2] Krizhevsky, Alex et al. “ImageNet classification with deep convolutional neural networks.” Communications of the ACM 60 (2012): 84 - 90.
[3] Simonyan, Karen and Andrew Zisserman. “Very Deep Convolutional Networks for Large-Scale Image Recognition.” CoRR abs/1409.1556 (2015): n. pag.
[4] Szegedy, Christian et al. “Going deeper with convolutions.” 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2015): 1-9.
[5] He, Kaiming et al. “Deep Residual Learning for Image Recognition.” 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2016): 770-778.
[6] Zeiler, Matthew D. and Rob Fergus. “Visualizing and Understanding Convolutional Networks.” ECCV (2014).