Reuse intermediate results in caffe

[JM]

Hi

I want to implement knowledge distillation in caffe. 

1. Are there any solution to reuse intermediate results to save GPU memory in caffe? 

2. How can I use different mode in the same net, like one branch for training and the other for testing(only to generate the final output)?

Thanks