def push(
self,
observation: np.array,
action: int,
next_observation: np.array,
reward: float,
done: bool,
):
""" Pushes a transition into the replay buffer. """
self.observations[self.position] = observation.clone()
self.actions[self.position] = torch.tensor([action], dtype=torch.long)
self.next_observations[self.position] = next_observation.clone()
self.rewards[self.position] = torch.tensor([reward], dtype=torch.float32)
self.dones[self.position] = torch.tensor([done], dtype=torch.bool)
self._update_stats(self.observations[self.position])
self.position = (self.position + 1) % len(self)
def xyxy_to_xywh(boxes_xyxy: np.array):
if isinstance(boxes_xyxy, torch.Tensor):
boxes_xywh = boxes_xyxy.clone()
elif isinstance(boxes_xyxy, np.ndarray):
boxes_xywh = boxes_xyxy.copy()
else:
raise TypeError
boxes_xywh[:, 0] = (boxes_xyxy[:, 0] + boxes_xyxy[:, 2]) / 2.
boxes_xywh[:, 1] = (boxes_xyxy[:, 1] + boxes_xyxy[:, 3]) / 2.
boxes_xywh[:, 2] = boxes_xyxy[:, 2] - boxes_xyxy[:, 0]
boxes_xywh[:, 3] = boxes_xyxy[:, 3] - boxes_xyxy[:, 1]
return boxes_xywh
