b900fdf6f2
Closes #947 This removes all kwargs from all policy constructors. While doing that, I also improved several names and added a whole lot of TODOs. ## Functional changes: 1. Added possibility to pass None as `critic2` and `critic2_optim`. In fact, the default behavior then should cover the absolute majority of cases 2. Added a function called `clone_optimizer` as a temporary measure to support passing `critic2_optim=None` ## Breaking changes: 1. `action_space` is no longer optional. In fact, it already was non-optional, as there was a ValueError in BasePolicy.init. So now several examples were fixed to reflect that 2. `reward_normalization` removed from DDPG and children. It was never allowed to pass it as `True` there, an error would have been raised in `compute_n_step_reward`. Now I removed it from the interface 3. renamed `critic1` and similar to `critic`, in order to have uniform interfaces. Note that the `critic` in DDPG was optional for the sole reason that child classes used `critic1`. I removed this optionality (DDPG can't do anything with `critic=None`) 4. Several renamings of fields (mostly private to public, so backwards compatible) ## Additional changes: 1. Removed type and default declaration from docstring. This kind of duplication is really not necessary 2. Policy constructors are now only called using named arguments, not a fragile mixture of positional and named as before 5. Minor beautifications in typing and code 6. Generally shortened docstrings and made them uniform across all policies (hopefully) ## Comment: With these changes, several problems in tianshou's inheritance hierarchy become more apparent. I tried highlighting them for future work. --------- Co-authored-by: Dominik Jain <d.jain@appliedai.de>
108 lines
4.1 KiB
Python
108 lines
4.1 KiB
Python
from typing import Any, cast
|
|
|
|
import numpy as np
|
|
import torch
|
|
|
|
from tianshou.data import ReplayBuffer, SegmentTree, to_numpy
|
|
from tianshou.data.types import PrioBatchProtocol, RolloutBatchProtocol
|
|
|
|
|
|
class PrioritizedReplayBuffer(ReplayBuffer):
|
|
"""Implementation of Prioritized Experience Replay. arXiv:1511.05952.
|
|
|
|
:param alpha: the prioritization exponent.
|
|
:param beta: the importance sample soft coefficient.
|
|
:param weight_norm: whether to normalize returned weights with the maximum
|
|
weight value within the batch. Default to True.
|
|
|
|
.. seealso::
|
|
|
|
Please refer to :class:`~tianshou.data.ReplayBuffer` for other APIs' usage.
|
|
"""
|
|
|
|
def __init__(
|
|
self,
|
|
size: int,
|
|
alpha: float,
|
|
beta: float,
|
|
weight_norm: bool = True,
|
|
**kwargs: Any,
|
|
) -> None:
|
|
# will raise KeyError in PrioritizedVectorReplayBuffer
|
|
# super().__init__(size, **kwargs)
|
|
ReplayBuffer.__init__(self, size, **kwargs)
|
|
assert alpha > 0.0
|
|
assert beta >= 0.0
|
|
self._alpha, self._beta = alpha, beta
|
|
self._max_prio = self._min_prio = 1.0
|
|
# save weight directly in this class instead of self._meta
|
|
self.weight = SegmentTree(size)
|
|
self.__eps = np.finfo(np.float32).eps.item()
|
|
self.options.update(alpha=alpha, beta=beta)
|
|
self._weight_norm = weight_norm
|
|
|
|
def init_weight(self, index: int | np.ndarray) -> None:
|
|
self.weight[index] = self._max_prio**self._alpha
|
|
|
|
def update(self, buffer: ReplayBuffer) -> np.ndarray:
|
|
indices = super().update(buffer)
|
|
self.init_weight(indices)
|
|
return indices
|
|
|
|
def add(
|
|
self,
|
|
batch: RolloutBatchProtocol,
|
|
buffer_ids: np.ndarray | list[int] | None = None,
|
|
) -> tuple[np.ndarray, np.ndarray, np.ndarray, np.ndarray]:
|
|
ptr, ep_rew, ep_len, ep_idx = super().add(batch, buffer_ids)
|
|
self.init_weight(ptr)
|
|
return ptr, ep_rew, ep_len, ep_idx
|
|
|
|
def sample_indices(self, batch_size: int) -> np.ndarray:
|
|
if batch_size > 0 and len(self) > 0:
|
|
scalar = np.random.rand(batch_size) * self.weight.reduce()
|
|
return self.weight.get_prefix_sum_idx(scalar) # type: ignore
|
|
return super().sample_indices(batch_size)
|
|
|
|
def get_weight(self, index: int | np.ndarray) -> float | np.ndarray:
|
|
"""Get the importance sampling weight.
|
|
|
|
The "weight" in the returned Batch is the weight on loss function to debias
|
|
the sampling process (some transition tuples are sampled more often so their
|
|
losses are weighted less).
|
|
"""
|
|
# important sampling weight calculation
|
|
# original formula: ((p_j/p_sum*N)**(-beta))/((p_min/p_sum*N)**(-beta))
|
|
# simplified formula: (p_j/p_min)**(-beta)
|
|
return (self.weight[index] / self._min_prio) ** (-self._beta)
|
|
|
|
def update_weight(self, index: np.ndarray, new_weight: np.ndarray | torch.Tensor) -> None:
|
|
"""Update priority weight by index in this buffer.
|
|
|
|
:param np.ndarray index: index you want to update weight.
|
|
:param np.ndarray new_weight: new priority weight you want to update.
|
|
"""
|
|
weight = np.abs(to_numpy(new_weight)) + self.__eps
|
|
self.weight[index] = weight**self._alpha
|
|
self._max_prio = max(self._max_prio, weight.max())
|
|
self._min_prio = min(self._min_prio, weight.min())
|
|
|
|
def __getitem__(self, index: slice | int | list[int] | np.ndarray) -> PrioBatchProtocol:
|
|
if isinstance(index, slice): # change slice to np array
|
|
# buffer[:] will get all available data
|
|
indices = (
|
|
self.sample_indices(0)
|
|
if index == slice(None)
|
|
else self._indices[: len(self)][index]
|
|
)
|
|
else:
|
|
indices = index # type: ignore
|
|
batch = super().__getitem__(indices)
|
|
weight = self.get_weight(indices)
|
|
# ref: https://github.com/Kaixhin/Rainbow/blob/master/memory.py L154
|
|
batch.weight = weight / np.max(weight) if self._weight_norm else weight
|
|
return cast(PrioBatchProtocol, batch)
|
|
|
|
def set_beta(self, beta: float) -> None:
|
|
self._beta = beta
|