(self, batch_size=32)
| 75 | self.size = min(self.size+1, self.max_size) |
| 76 | |
| 77 | def sample_batch(self, batch_size=32): |
| 78 | idxs = np.random.randint(0, self.size, size=batch_size) |
| 79 | return dict(s=self.obs1_buf[idxs], |
| 80 | s2=self.obs2_buf[idxs], |
| 81 | a=self.acts_buf[idxs], |
| 82 | r=self.rews_buf[idxs], |
| 83 | d=self.done_buf[idxs]) |
| 84 | |
| 85 | |
| 86 | ### Implement the DDPG algorithm ### |