![](/img/trans.png)
[英]Keras: clean implementation for multiple outputs and custom loss functions?
[英]Keras - Implementation of custom loss function with multiple outputs
我正在尝试复制(一个更小的版本)AlphaGo Zero 系统。 但是,在网络 model 中,我遇到了问题。 我应该实现的损失 function 如下:
在哪里:
我向网络传递一个通道列表(表示游戏状态)和一个数组(大小相同的pi和p ),表示哪些动作确实有效(如果有效则输入1
,否则输入0
)。
如您所见,损失 function 使用目标和网络预测进行计算。 但是经过广泛的搜索,在实现我的自定义损失 function 时,即使我有两个“y_true”和两个“y_pred”,我也只能作为参数y_true
和y_pred
传递。 我曾尝试使用索引来获取这些值,但我很确定它不起作用。
网络的建模和自定义损失 function 在下面的代码中:
def custom_loss(y_true, y_pred):
# I am pretty sure this does not work
output_prob_dist = y_pred[0]
output_value = y_pred[1]
label_prob_dist = y_true[0]
label_value = y_pred[1]
mse_loss = K.mean(K.square(label_value - output_value), axis=-1)
cross_entropy_loss = K.dot(K.transpose(label_prob_dist), output_prob_dist)
return mse_loss - cross_entropy_loss
def define_model():
"""Neural Network model implementation using Keras + Tensorflow."""
state_channels = Input(shape = (5,5,6), name='States_Channels_Input')
valid_actions_dist = Input(shape = (32,), name='Valid_Actions_Input')
conv = Conv2D(filters=10, kernel_size=2, kernel_regularizer=regularizers.l2(0.0001), activation='relu', name='Conv_Layer')(state_channels)
pool = MaxPooling2D(pool_size=(2, 2), name='Pooling_Layer')(conv)
flat = Flatten(name='Flatten_Layer')(pool)
# Merge of the flattened channels (after pooling) and the valid action
# distribution. Used only as input in the probability distribution head.
merge = concatenate([flat, valid_actions_dist])
#Probability distribution over actions
hidden_fc_prob_dist_1 = Dense(100, kernel_regularizer=regularizers.l2(0.0001), activation='relu', name='FC_Prob_1')(merge)
hidden_fc_prob_dist_2 = Dense(100, kernel_regularizer=regularizers.l2(0.0001), activation='relu', name='FC_Prob_2')(hidden_fc_prob_dist_1)
output_prob_dist = Dense(32, kernel_regularizer=regularizers.l2(0.0001), activation='softmax', name='Output_Dist')(hidden_fc_prob_dist_2)
#Value of a state
hidden_fc_value_1 = Dense(100, kernel_regularizer=regularizers.l2(0.0001), activation='relu', name='FC_Value_1')(flat)
hidden_fc_value_2 = Dense(100, kernel_regularizer=regularizers.l2(0.0001), activation='relu', name='FC_Value_2')(hidden_fc_value_1)
output_value = Dense(1, kernel_regularizer=regularizers.l2(0.0001), activation='tanh', name='Output_Value')(hidden_fc_value_2)
model = Model(inputs=[state_channels, valid_actions_dist], outputs=[output_prob_dist, output_value])
model.compile(loss=custom_loss, optimizer='adam', metrics=['accuracy'])
return model
# In the main method
model = define_model()
# ...
# MCTS routine to collect the data for the network input
# ...
x_train = [channels_input, valid_actions_dist_input]
y_train = [dist_probs_label, who_won_label]
model.fit(x_train, y_train, epochs=10)
简而言之,我的问题是:如何正确实现此自定义损失 function 使用网络输出和网络的 label 值?
我检查了他们的 git 并且发生了很多事情; 如方程式所示,最终损失是三个不同损失的组合,三个网络正在最小化这个最终损失。 他们的损失代码如下:
# train ops
policy_cost = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits_v2(
logits=logits, labels=tf.stop_gradient(labels['pi_tensor'])))
value_cost = params['value_cost_weight'] * tf.reduce_mean(
tf.square(value_output - labels['value_tensor']))
reg_vars = [v for v in tf.trainable_variables()
if 'bias' not in v.name and 'beta' not in v.name]
l2_cost = params['l2_strength'] * \
tf.add_n([tf.nn.l2_loss(v) for v in reg_vars])
combined_cost = policy_cost + value_cost + l2_cost
您可以参考此内容并相应地进行更改。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.