(self, batch_size=32)
| 44 | self.size = min(self.size+1, self.max_size) |
| 45 | |
| 46 | def sample_batch(self, batch_size=32): |
| 47 | idxs = np.random.randint(0, self.size, size=batch_size) |
| 48 | return dict(s=self.obs1_buf[idxs], |
| 49 | s2=self.obs2_buf[idxs], |
| 50 | a=self.acts_buf[idxs], |
| 51 | r=self.rews_buf[idxs], |
| 52 | d=self.done_buf[idxs]) |
| 53 | |
| 54 | |
| 55 |