Method forward

rl2/mountaincar/pg_tf.py:41–46

(self, X)

    self.f = f

  def forward(self, X):
    if self.use_bias:
      a = tf.matmul(X, self.W) + self.b
    else:
      a = tf.matmul(X, self.W)
    return self.f(a)


# approximates pi(a | s)