(self, state)
| 13 | self.max_action = max_action |
| 14 | |
| 15 | def forward(self, state): |
| 16 | a = F.relu(self.l1(state)) |
| 17 | a = F.relu(self.l2(a)) |
| 18 | |
| 19 | return self.max_action * F.tanh(self.l3(a)) |
| 20 | |
| 21 | def select_action(self, state): |
| 22 | state = paddle.to_tensor(state.reshape(1, -1)).astype('float32') |