{# Copyright (c) 2014-2015, NVIDIA CORPORATION. All rights reserved. #} {% from "helper.html" import print_flashes %} {% extends "layout.html" %} {% block title %} New Image Classification Model {% endblock %} {% block nav %}
Show advanced learning rate options
In this mode, you must specify 1 (or 2) Data layers, an Accuracy layer, and a SoftmaxWithLoss layer. During deployment, the SoftmaxWithLoss layer will be replaced with a Softmax layer.
Once Caffe implements their "all-in-one nets" feature to consolidate train_val and deploy nets, this will become much more extensible.