Pownie/backgammon
1
0
Fork 0
Code Issues 6 Pull Requests Packages Projects Releases Wiki Activity

Custom activation (2*tanh(x)) function implemented with tensorflow primitives.

Browse Source
This commit is contained in:
Anders B. Ladefoged 2018-03-06 12:19:04 +01:00
parent 5845edf084
commit c9e4446a52
1 changed files with 6 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
network.py
View File

@ -19,6 +19,10 @@ class Network:
# TODO: Actually compile tensorflow properly # TODO: Actually compile tensorflow properly
os.environ["TF_CPP_MIN_LOG_LEVEL"]="2" os.environ["TF_CPP_MIN_LOG_LEVEL"]="2"
def custom_tanh(self, x, name=None):
a = tf.Variable(2.00, tf.float32)
return tf.scalar_mul(a, tf.tanh(x, name))
def __init__(self, session): def __init__(self, session):
self.session = session self.session = session
self.config = Config self.config = Config
@ -40,12 +44,10 @@ class Network:
b_1 = tf.zeros(hidden_size,) b_1 = tf.zeros(hidden_size,)
b_2 = tf.zeros(output_size,) b_2 = tf.zeros(output_size,)
value_after_input = tf.sigmoid(tf.matmul(self.x, W_1) + b_1, name='hidden_layer') value_after_input = self.custom_tanh(tf.matmul(self.x, W_1) + b_1, name='hidden_layer')
# TODO: Remember to make this tanh * 2 # TODO: Remember to make this tanh * 2
# self.value = tf.layers.dense(input=value_after_input, units=hidden_size, \ self.value = self.custom_tanh(tf.matmul(value_after_input, W_2) + b_2, name='output_layer')
# activation=self.custom_tanh, kernel_initializer=xavier_init())
self.value = 2*tf.nn.tanh(tf.matmul(value_after_input, W_2) + b_2, name='output_layer')
# tf.reduce_sum basically finds the sum of it's input, so this gives the difference between the two values, in case they should be lists, which they might be if our input changes # tf.reduce_sum basically finds the sum of it's input, so this gives the difference between the two values, in case they should be lists, which they might be if our input changes
difference_in_values = tf.reduce_sum(self.value_next - self.value, name='difference') difference_in_values = tf.reduce_sum(self.value_next - self.value, name='difference')
@ -67,7 +69,6 @@ class Network:
self.saver = tf.train.Saver(max_to_keep=1) self.saver = tf.train.Saver(max_to_keep=1)
self.session.run(tf.global_variables_initializer()) self.session.run(tf.global_variables_initializer())
def eval_state(self, state): def eval_state(self, state):
# Run state through a network # Run state through a network