# Copyright (c) 2017 PaddlePaddle Authors. All Rights Reserved
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

from paddle.trainer_config_helpers import *

################################### Data Configuration ###################################
TrainData(ProtoData(files = "trainer/tests/mnist.list"))
################################### Algorithm Configuration ###################################
settings(batch_size = 256,
         learning_method = MomentumOptimizer(momentum=0.5, sparse=False))
################################### Network Configuration ###################################
data = data_layer(name ="input", size=784)

tmp = img_conv_layer(input=data,
            num_channels=1,
            filter_size=3,
            num_filters=32,
            padding=1,
            shared_biases=True,
            act=ReluActivation())

a1 = img_conv_layer(input=tmp,
            filter_size=1,
            num_filters=32,
            padding=0,
            shared_biases=True,
            act=ReluActivation())

a2 = img_conv_layer(input=tmp,
            filter_size=3,
            num_filters=32,
            padding=1,
            shared_biases=True,
            act=ReluActivation())

tmp = concat_layer(input=[a1, a2])

tmp = img_pool_layer(input=tmp,
            num_channels=64,
            pool_size=3,
            stride=2,
            padding=1,
            pool_type=AvgPooling())

b1 = img_conv_layer(input=tmp,
            filter_size=3,
            num_filters=64,
            padding=1,
            shared_biases=True,
            act=ReluActivation())

b1 = img_pool_layer(input=b1,
            pool_size=3,
            stride=1,
            padding=1,
            pool_type=MaxPooling())

b2 = img_conv_layer(input=tmp,
            filter_size=5,
            num_filters=64,
            padding=2,
            shared_biases=True,
            act=ReluActivation())

b2 = img_pool_layer(input=b2,
            pool_size=5,
            stride=1,
            padding=2,
            pool_type=MaxPooling())

tmp = addto_layer(input=[b1, b2],
            act=ReluActivation(),
            bias_attr=False)

tmp = img_pool_layer(input=tmp,
            pool_size=3,
            stride=2,
            padding=1,
            pool_type=MaxPooling())

tmp = fc_layer(input=tmp, size=64,
            bias_attr=False,
            act=TanhActivation())

output = fc_layer(input=tmp, size=10,
            bias_attr=True,
            act=SoftmaxActivation())

lbl = data_layer(name ="label", size=10)

cost = classification_cost(input=output, label=lbl)
outputs(cost)