D4RLExperienceReplay¶
- class torchrl.data.datasets.D4RLExperienceReplay(dataset_id, batch_size: int, sampler: Sampler | None = None, writer: Writer | None = None, collate_fn: Callable | None = None, pin_memory: bool = False, prefetch: int | None = None, transform: 'torchrl.envs.Transform' | None = None, split_trajs: bool = False, from_env: bool = False, use_truncated_as_done: bool = True, direct_download: bool = None, terminate_on_end: bool = None, download: bool = True, root: str | Path | None = None, **env_kwargs)[source]¶
D4RL 的经验回放类。
要安装 D4RL,请按照官方仓库中的说明进行操作。
数据格式遵循TED 约定。回放缓冲区包含 D4RLExperienceReplay.specs 下的环境规范。
如果存在,元数据将写入
D4RLExperienceReplay.metadata
并从数据集中排除。转换使用
done = terminated | truncated
重建,并且"done"
状态的("next", "observation")
为零。- 参数:
dataset_id (str) – 要从中获取数据的 D4RL 环境的 dataset_id。
batch_size (int) – 采样期间使用的批次大小。
sampler (Sampler, 可选) – 要使用的采样器。如果未提供,将使用默认的 RandomSampler()。
writer (Writer, 可选) – 要使用的写入器。如果未提供,将使用默认的
ImmutableDatasetWriter
。collate_fn (callable, 可选) – 合并样本列表以形成张量/输出的小批量。在使用来自映射样式数据集的批量加载时使用。
pin_memory (bool) – 是否应在 rb 样本上调用 pin_memory()。
prefetch (int, 可选) – 使用多线程预取的下一个批次的数目。
transform (Transform, 可选) – 调用 sample() 时要执行的转换。要链接转换,请使用
Compose
类。split_trajs (bool, optional) – 如果
True
,轨迹将沿第一个维度分割并填充以具有匹配的形状。为了分割轨迹,将使用"done"
信号,该信号通过done = truncated | terminated
恢复。换句话说,假设任何truncated
或terminated
信号等效于轨迹的结束。对于来自D4RL
的某些数据集,这可能不正确。用户需要根据split_trajs
的这种用法做出准确的选择。默认为False
。from_env (bool, optional) –
如果
True
,将使用env.get_dataset()
检索数据集。否则将使用d4rl.qlearning_dataset()
。默认为True
。注意
使用
from_env=False
将提供比from_env=True
更少的数据。例如,信息键将被省略。通常,from_env=False
且terminate_on_end=True
将导致与from_env=True
相同的结果,后者包含前者不具有的元数据和信息条目。注意
from_env=True
和from_env=False
中的键 *可能* 会意外地不同。特别是,"truncated"
键(用于确定一个episode的结束)在from_env=False
时可能不存在,但在其他情况下存在,导致在启用traj_splits
时切片不同。direct_download (bool) – 如果
True
,数据将被下载而无需 D4RL。如果为None
,如果环境中存在d4rl
,它将用于下载数据集,否则下载将回退到direct_download=True
。这与from_env=True
不兼容。默认为None
。use_truncated_as_done (bool, optional) – 如果
True
,done = terminated | truncated
。否则,仅使用terminated
键。默认为True
。terminate_on_end (bool, optional) – 在轨迹的最后一个时间步长设置
done=True
。默认为False
,并将丢弃每个轨迹的最后一个时间步长。这仅与direct_download=False
一起使用。root (Path 或 str, optional) – D4RL 数据集根目录。实际的数据集内存映射文件将保存在 <root>/<dataset_id> 下。如果未提供,则默认为 ``~/.cache/torchrl/d4rl`。
download (bool, optional) – 数据集是否应在找不到时下载。默认为
True
。**env_kwargs (键值对) –
d4rl.qlearning_dataset()
的其他关键字参数。
示例
>>> from torchrl.data.datasets.d4rl import D4RLExperienceReplay >>> from torchrl.envs import ObservationNorm >>> data = D4RLExperienceReplay("maze2d-umaze-v1", 128) >>> # we can append transforms to the dataset >>> data.append_transform(ObservationNorm(loc=-1, scale=1.0, in_keys=["observation"])) >>> data.sample(128)
- add(data: TensorDictBase) int ¶
将单个元素添加到回放缓冲区。
- 参数:
data (Any) – 要添加到回放缓冲区的数据
- 返回:
数据在回放缓冲区中的索引。
- append_transform(transform: Transform, *, invert: bool = False) ReplayBuffer ¶
在末尾追加转换。
当调用 sample 时,转换将按顺序应用。
- 参数:
transform (Transform) – 要追加的转换
- 关键字参数:
invert (bool, optional) – 如果
True
,转换将被反转(正向调用将在写入期间调用,反向调用将在读取期间调用)。默认为False
。
示例
>>> rb = ReplayBuffer(storage=LazyMemmapStorage(10), batch_size=4) >>> data = TensorDict({"a": torch.zeros(10)}, [10]) >>> def t(data): ... data += 1 ... return data >>> rb.append_transform(t, invert=True) >>> rb.extend(data) >>> assert (data == 1).all()
- property data_path: Path¶
数据集的路径,包括分割。
- property data_path_root: Path¶
数据集根目录的路径。
- delete()¶
从磁盘删除数据集存储。
- dumps(path)¶
将回放缓冲区保存到指定路径的磁盘上。
- 参数:
path (Path 或 str) – 保存回放缓冲区的路径。
示例
>>> import tempfile >>> import tqdm >>> from torchrl.data import LazyMemmapStorage, TensorDictReplayBuffer >>> from torchrl.data.replay_buffers.samplers import PrioritizedSampler, RandomSampler >>> import torch >>> from tensordict import TensorDict >>> # Build and populate the replay buffer >>> S = 1_000_000 >>> sampler = PrioritizedSampler(S, 1.1, 1.0) >>> # sampler = RandomSampler() >>> storage = LazyMemmapStorage(S) >>> rb = TensorDictReplayBuffer(storage=storage, sampler=sampler) >>> >>> for _ in tqdm.tqdm(range(100)): ... td = TensorDict({"obs": torch.randn(100, 3, 4), "next": {"obs": torch.randn(100, 3, 4)}, "td_error": torch.rand(100)}, [100]) ... rb.extend(td) ... sample = rb.sample(32) ... rb.update_tensordict_priority(sample) >>> # save and load the buffer >>> with tempfile.TemporaryDirectory() as tmpdir: ... rb.dumps(tmpdir) ... ... sampler = PrioritizedSampler(S, 1.1, 1.0) ... # sampler = RandomSampler() ... storage = LazyMemmapStorage(S) ... rb_load = TensorDictReplayBuffer(storage=storage, sampler=sampler) ... rb_load.loads(tmpdir) ... assert len(rb) == len(rb_load)
- empty()¶
清空回放缓冲区并将光标重置为 0。
- extend(tensordicts: TensorDictBase) Tensor ¶
使用可迭代对象中包含的一个或多个元素扩展回放缓冲区。
如果存在,将调用反向转换。
- 参数:
data (iterable) – 要添加到回放缓冲区的数据集合。
- 返回:
添加到回放缓冲区的数据的索引。
警告
当处理值列表时,
extend()
可能具有模棱两可的签名,这些值应被解释为 PyTree(在这种情况下,列表中的所有元素都将被放入存储中存储的 PyTree 中的一个切片)或要逐个添加的值列表。为了解决这个问题,TorchRL 在列表和元组之间做出了明确的区别:元组将被视为 PyTree,列表(在根级别)将被解释为要逐个添加到缓冲区的值的堆栈。对于ListStorage
实例,只能提供未绑定的元素(没有 PyTree)。
- insert_transform(index: int, transform: Transform, *, invert: bool = False) ReplayBuffer ¶
插入转换。
当调用 sample 时,转换将按顺序执行。
- 参数:
index (int) – 插入转换的位置。
transform (Transform) – 要追加的转换
- 关键字参数:
invert (bool, optional) – 如果
True
,转换将被反转(正向调用将在写入期间调用,反向调用将在读取期间调用)。默认为False
。
- loads(path)¶
加载给定路径下的回放缓冲区状态。
缓冲区应具有匹配的组件,并使用
dumps()
保存。- 参数:
path (Path 或 str) – 重放缓冲区保存的路径。
有关更多信息,请参阅
dumps()
。
- preprocess(fn: Callable[[TensorDictBase], TensorDictBase], dim: int = 0, num_workers: int | None = None, *, chunksize: int | None = None, num_chunks: int | None = None, pool: mp.Pool | None = None, generator: torch.Generator | None = None, max_tasks_per_child: int | None = None, worker_threads: int = 1, index_with_generator: bool = False, pbar: bool = False, mp_start_method: str | None = None, num_frames: int | None = None, dest: str | Path) TensorStorage ¶
预处理数据集并返回一个包含格式化数据的新的存储。
数据转换必须是单一的(作用于数据集的单个样本)。
参数和关键字参数被转发到
map()
。随后可以使用
delete()
删除数据集。- 关键字参数:
dest (path 或 等效项) – 新数据集位置的路径。
num_frames (int, 可选) – 如果提供,则仅转换前num_frames。这对于首先调试转换很有用。
返回:一个新的存储,可在
ReplayBuffer
实例中使用。示例
>>> from torchrl.data.datasets import MinariExperienceReplay >>> >>> data = MinariExperienceReplay( ... list(MinariExperienceReplay.available_datasets)[0], ... batch_size=32 ... ) >>> print(data) MinariExperienceReplay( storages=TensorStorage(TensorDict( fields={ action: MemoryMappedTensor(shape=torch.Size([1000000, 8]), device=cpu, dtype=torch.float32, is_shared=True), episode: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.int64, is_shared=True), info: TensorDict( fields={ distance_from_origin: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), forward_reward: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), goal: MemoryMappedTensor(shape=torch.Size([1000000, 2]), device=cpu, dtype=torch.float64, is_shared=True), qpos: MemoryMappedTensor(shape=torch.Size([1000000, 15]), device=cpu, dtype=torch.float64, is_shared=True), qvel: MemoryMappedTensor(shape=torch.Size([1000000, 14]), device=cpu, dtype=torch.float64, is_shared=True), reward_ctrl: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), reward_forward: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), reward_survive: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), success: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.bool, is_shared=True), x_position: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), x_velocity: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), y_position: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), y_velocity: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True)}, batch_size=torch.Size([1000000]), device=cpu, is_shared=False), next: TensorDict( fields={ done: MemoryMappedTensor(shape=torch.Size([1000000, 1]), device=cpu, dtype=torch.bool, is_shared=True), info: TensorDict( fields={ distance_from_origin: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), forward_reward: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), goal: MemoryMappedTensor(shape=torch.Size([1000000, 2]), device=cpu, dtype=torch.float64, is_shared=True), qpos: MemoryMappedTensor(shape=torch.Size([1000000, 15]), device=cpu, dtype=torch.float64, is_shared=True), qvel: MemoryMappedTensor(shape=torch.Size([1000000, 14]), device=cpu, dtype=torch.float64, is_shared=True), reward_ctrl: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), reward_forward: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), reward_survive: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), success: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.bool, is_shared=True), x_position: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), x_velocity: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), y_position: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True), y_velocity: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.float64, is_shared=True)}, batch_size=torch.Size([1000000]), device=cpu, is_shared=False), observation: TensorDict( fields={ achieved_goal: MemoryMappedTensor(shape=torch.Size([1000000, 2]), device=cpu, dtype=torch.float64, is_shared=True), desired_goal: MemoryMappedTensor(shape=torch.Size([1000000, 2]), device=cpu, dtype=torch.float64, is_shared=True), observation: MemoryMappedTensor(shape=torch.Size([1000000, 27]), device=cpu, dtype=torch.float64, is_shared=True)}, batch_size=torch.Size([1000000]), device=cpu, is_shared=False), reward: MemoryMappedTensor(shape=torch.Size([1000000, 1]), device=cpu, dtype=torch.float64, is_shared=True), terminated: MemoryMappedTensor(shape=torch.Size([1000000, 1]), device=cpu, dtype=torch.bool, is_shared=True), truncated: MemoryMappedTensor(shape=torch.Size([1000000, 1]), device=cpu, dtype=torch.bool, is_shared=True)}, batch_size=torch.Size([1000000]), device=cpu, is_shared=False), observation: TensorDict( fields={ achieved_goal: MemoryMappedTensor(shape=torch.Size([1000000, 2]), device=cpu, dtype=torch.float64, is_shared=True), desired_goal: MemoryMappedTensor(shape=torch.Size([1000000, 2]), device=cpu, dtype=torch.float64, is_shared=True), observation: MemoryMappedTensor(shape=torch.Size([1000000, 27]), device=cpu, dtype=torch.float64, is_shared=True)}, batch_size=torch.Size([1000000]), device=cpu, is_shared=False)}, batch_size=torch.Size([1000000]), device=cpu, is_shared=False)), samplers=RandomSampler, writers=ImmutableDatasetWriter(), batch_size=32, transform=Compose( ), collate_fn=<function _collate_id at 0x120e21dc0>) >>> from torchrl.envs import CatTensors, Compose >>> from tempfile import TemporaryDirectory >>> >>> cat_tensors = CatTensors( ... in_keys=[("observation", "observation"), ("observation", "achieved_goal"), ... ("observation", "desired_goal")], ... out_key="obs" ... ) >>> cat_next_tensors = CatTensors( ... in_keys=[("next", "observation", "observation"), ... ("next", "observation", "achieved_goal"), ... ("next", "observation", "desired_goal")], ... out_key=("next", "obs") ... ) >>> t = Compose(cat_tensors, cat_next_tensors) >>> >>> def func(td): ... td = td.select( ... "action", ... "episode", ... ("next", "done"), ... ("next", "observation"), ... ("next", "reward"), ... ("next", "terminated"), ... ("next", "truncated"), ... "observation" ... ) ... td = t(td) ... return td >>> with TemporaryDirectory() as tmpdir: ... new_storage = data.preprocess(func, num_workers=4, pbar=True, mp_start_method="fork", dest=tmpdir) ... rb = ReplayBuffer(storage=new_storage) ... print(rb) ReplayBuffer( storage=TensorStorage( data=TensorDict( fields={ action: MemoryMappedTensor(shape=torch.Size([1000000, 8]), device=cpu, dtype=torch.float32, is_shared=True), episode: MemoryMappedTensor(shape=torch.Size([1000000]), device=cpu, dtype=torch.int64, is_shared=True), next: TensorDict( fields={ done: MemoryMappedTensor(shape=torch.Size([1000000, 1]), device=cpu, dtype=torch.bool, is_shared=True), obs: MemoryMappedTensor(shape=torch.Size([1000000, 31]), device=cpu, dtype=torch.float64, is_shared=True), observation: TensorDict( fields={ }, batch_size=torch.Size([1000000]), device=cpu, is_shared=False), reward: MemoryMappedTensor(shape=torch.Size([1000000, 1]), device=cpu, dtype=torch.float64, is_shared=True), terminated: MemoryMappedTensor(shape=torch.Size([1000000, 1]), device=cpu, dtype=torch.bool, is_shared=True), truncated: MemoryMappedTensor(shape=torch.Size([1000000, 1]), device=cpu, dtype=torch.bool, is_shared=True)}, batch_size=torch.Size([1000000]), device=cpu, is_shared=False), obs: MemoryMappedTensor(shape=torch.Size([1000000, 31]), device=cpu, dtype=torch.float64, is_shared=True), observation: TensorDict( fields={ }, batch_size=torch.Size([1000000]), device=cpu, is_shared=False)}, batch_size=torch.Size([1000000]), device=cpu, is_shared=False), shape=torch.Size([1000000]), len=1000000, max_size=1000000), sampler=RandomSampler(), writer=RoundRobinWriter(cursor=0, full_storage=True), batch_size=None, collate_fn=<function _collate_id at 0x168406fc0>)
- register_load_hook(hook: Callable[[Any], Any])¶
注册存储的加载钩子。
注意
当前在保存重放缓冲区时不会序列化钩子:每次创建缓冲区时都必须手动重新初始化它们。
- register_save_hook(hook: Callable[[Any], Any])¶
注册存储的保存钩子。
注意
当前在保存重放缓冲区时不会序列化钩子:每次创建缓冲区时都必须手动重新初始化它们。
- sample(batch_size: int | None = None, return_info: bool = False, include_info: bool = None) TensorDictBase ¶
从重放缓冲区采样一批数据。
使用 Sampler 采样索引,并从 Storage 中检索它们。
- 参数:
batch_size (int, 可选) – 要收集的数据的大小。如果没有提供,此方法将根据采样器指示采样批次大小。
return_info (bool) – 是否返回信息。如果为 True,则结果为元组 (data, info)。如果为 False,则结果为数据。
- 返回:
包含在重放缓冲区中选择的批次数据的张量字典。如果 return_info 标志设置为 True,则包含此张量字典和信息的元组。