55bab51a583edf62328c654f72321d3f0caf62a2,yarll/agents/sac.py,SAC,build_target_value_network,#SAC#Any#,152

Before Change



        x = self.states
        x = tf.nn.relu(tf.nn.xw_plus_b(x, target_net[0], target_net[1]))
        x = tf.nn.relu(tf.nn.xw_plus_b(x, target_net[2], target_net[3]))
        value = tf.nn.xw_plus_b(x, target_net[4], target_net[5], name="target_value")

        return value, target_update

After Change


        target_net = [ema.average(v) for v in value_vars]

        x = self.states
        for i in range(self.config["n_hidden_layers"]):
            x = tf.nn.relu(tf.nn.xw_plus_b(x, target_net[i * 2], target_net[i * 2 + 1]))
        value = tf.nn.xw_plus_b(x, target_net[i + 1], target_net[i + 2], name="target_value")
Italian Trulli
In pattern: SUPERPATTERN

Frequency: 3

Non-data size: 4

Instances


Project Name: arnomoonens/yarll
Commit Name: 55bab51a583edf62328c654f72321d3f0caf62a2
Time: 2018-11-14
Author: arno.moonens@gmail.com
File Name: yarll/agents/sac.py
Class Name: SAC
Method Name: build_target_value_network


Project Name: pyprob/pyprob
Commit Name: d36137596ae67d1b07138deca4a42fbbf4887b7e
Time: 2018-09-19
Author: atilimgunes.baydin@gmail.com
File Name: pyprob/nn/embedding_feedforward.py
Class Name: EmbeddingFeedForward
Method Name: forward


Project Name: stanfordnlp/stanza
Commit Name: 09d7bda36be8fe7ff22bc9f10e9650b12326c8ce
Time: 2018-06-30
Author: qipeng@noreply.github.com
File Name: models/tokenize/tokenizer_models.py
Class Name: RNNTokenizer
Method Name: forward