(self, batch_size=32)
| 52 | self.size = min(self.size+1, self.max_size) |
| 53 | |
| 54 | def sample_batch(self, batch_size=32): |
| 55 | idxs = np.random.randint(0, self.size, size=batch_size) |
| 56 | return dict(s=self.obs1_buf[idxs], |
| 57 | s2=self.obs2_buf[idxs], |
| 58 | a=self.acts_buf[idxs], |
| 59 | r=self.rews_buf[idxs], |
| 60 | d=self.done_buf[idxs]) |
| 61 | |
| 62 | |
| 63 |