(self, s, a)
| 62 | return [self.predict(s, a) for a in range(self.env.action_space.n)] |
| 63 | |
| 64 | def grad(self, s, a): |
| 65 | sa = np.concatenate((s, [a])) |
| 66 | x = self.featurizer.transform([sa])[0] |
| 67 | return x |
| 68 | |
| 69 | |
| 70 | def test_agent(model, env, n_episodes=20): |