Machine learning 在一行中添加相同的CONV2d图层与继续添加过滤器大小增加的图层相比是否有好处
在一行中添加相同的图层是否有如下好处:Machine learning 在一行中添加相同的CONV2d图层与继续添加过滤器大小增加的图层相比是否有好处,machine-learning,keras,deep-learning,computer-vision,Machine Learning,Keras,Deep Learning,Computer Vision,在一行中添加相同的图层是否有如下好处: model.add(layers.Conv2D(32, (3, 3), activation='relu')) model.add(layers.Conv2D(32, (3, 3), activation='relu')) model.add(layers.Conv2D(32, (3, 3), activation='relu')) model.add(layers.Conv2D(64, (3, 3), activation='relu')) model
model.add(layers.Conv2D(32, (3, 3), activation='relu'))
model.add(layers.Conv2D(32, (3, 3), activation='relu'))
model.add(layers.Conv2D(32, (3, 3), activation='relu'))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.Conv2D(32, (3, 3), activation='relu'))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
与继续增加过滤器大小不同,如下所示:
model.add(layers.Conv2D(32, (3, 3), activation='relu'))
model.add(layers.Conv2D(32, (3, 3), activation='relu'))
model.add(layers.Conv2D(32, (3, 3), activation='relu'))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.Conv2D(32, (3, 3), activation='relu'))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
添加相同的层是否常见,因为继续增加过滤器的数量成本更高
当我在一行中使用相同的图层时,我获得了更高的准确性,但当我浏览示例和书籍时,我很少遇到添加相同图层的作者,我只是好奇为什么。这个问题很好,但答案取决于具体情况
谢谢你,Timbus!