{# Copyright (c) 2015-2016, NVIDIA CORPORATION. All rights reserved. #}
Use this field to enter a Caffe network in .prototxt format.
Classification are special in a few ways:
The network should include the necessary layers for all computation states - Train, Val and Deploy. When searching for layers, the network will be set to one of these three states:
Your layer's include and exclude rules should be set properly to be picked up by the desired computation state[s]. You can look at the standard networks for examples. In addition, since there are many ways to specify the rules, here's a table that you may find useful:
Rule | Train | Val | Deploy |
---|---|---|---|
none | |||
include { phase: TRAIN } | |||
include { stage: "train" } | |||
include { phase: TEST } | |||
include { stage: "val" } | |||
include { stage: "deploy" } | |||
exclude { stage: "deploy" } | |||
include { stage: "train" } include { stage: "val" } |
You may notice that DIGITS still generates separate files for train_val and deploy networks, and that they each only use phase and not stage. This is because Caffe doesn't fully support all-in-one networks yet.
Use this field to enter a Torch network using Lua code. Refer to the documentation for more information.