(self, state, sess)
| 176 | return actions[0] |
| 177 | |
| 178 | def get_value_prediction(self, state, sess): |
| 179 | # Make input N x D (N = 1) |
| 180 | feed_dict = { self.value_net.states: [state] } |
| 181 | vhat = sess.run(self.value_net.vhat, feed_dict) |
| 182 | # Prediction is a 1-D array of length N, just want the first value |
| 183 | return vhat[0] |
| 184 | |
| 185 | def run_n_steps(self, n, sess): |
| 186 | steps = [] |