Caffe is the most widely used deep learning platform in the field of image at present, and many Daniel at home and abroad use Caffe as the experimental platform. Personally, there are two main reasons. First, caffe developed earlier, and accumulated many research results of deep learning over time, such as code that can be run directly and pre-trained model, which can be easily used in experiments; Second, if future generations want to compare with the previous methods, they need to keep other factors consistent except the methods, such as the data used and the platform used in the experiment.

Vgg- 16 is a deep convolution neural network model, and 16 represents its depth. It is one of the representative depth models after alexnet, and has achieved good results in image classification and other tasks.

Googlenet is a deep convolution neural network model designed by Google. The first edition is 22 stories deep. The network adopts the idea of sparse learning, and the network scale is increased by sparse network parameters.

Caffe-like platforms include tensorflow, theano, torch, paddle and so on.

Vgg- 16, googlenet and other network structures are diverse, and the representative ones are alexnet and resnet.