Meng, Peng

For prediction, we don't need dropout. If set dropout to 1. The right behavior is dropout layer return directly.

We can create two Unet with different keep_prob for training and validation. How do you think about it? Since the Dropout layer is very time-consuming, it is better to skip...

For inference, the Dropout is about 16% of iteration time. Second row of this picture. We don't need to train two model. Just need to create a new model (model...

Hi @jakeret , any comment on the data. The data is based on CPU.

I am sorry for reply later. How about input two nets when creating Trainer object. The train_net for train, and the validation_net for validation. train_net can save the model for...

There is no weight for the dropout layer, it is ok to save model in the train net, and restore them in the validation net.

Fix this by change: src/caffe/test/test_layer_factory.cpp:line 33 if (iter->first == "Data") { to if(iter->first == "Data" || iter->first == "BoxData")