GenDGRLExperienceReplay¶
- class torchrl.data.datasets.GenDGRLExperienceReplay(dataset_id: str, batch_size: int = None, *, download: bool = True, root: str | None = None, **kwargs)[原始碼]¶
Gen-DGRL 經驗回放資料集。
此資料集配套論文“離線強化學習中的泛化差距”提供。
Arxiv: https://arxiv.org/abs/2312.05742
GitHub: https://github.com/facebookresearch/gen_dgrl
資料格式遵循 TED 規範。
此類提供對 ProcGen 資料集的訪問。在 GenDGRLExperienceReplay.available_datasets 中註冊的每個 dataset_id 包含一個特定任務(“bigfish”、“bossfight” 等),該任務透過逗號(“bigfish-1M_E” 等)與一個類別(“1M_E”、“1M_S” 等)分開。
在下載和準備過程中,資料以下載為 .tar 檔案,其中每個軌跡都獨立儲存在一個 .npy 檔案中。這些檔案會被提取,寫入連續的 mmap 張量,然後清除。此過程每個資料集可能需要幾分鐘。在叢集上,建議首先在不同的 worker 或程序上單獨對不同資料集執行下載和預處理,然後再啟動訓練指令碼。
- 引數:
dataset_id (str) – 要下載的資料集。必須是
GenDGRLExperienceReplay.available_datasets的一部分。batch_size (int, optional) – 取樣期間使用的批次大小。如果需要,可以透過 data.sample(batch_size) 覆蓋此值。
- 關鍵字引數:
root (Path or str, optional) –
GenDGRLExperienceReplay資料集的根目錄。實際的資料集記憶體對映檔案將儲存在 <root>/<dataset_id> 下。如果未提供,則預設為 ~/.cache/torchrl/atari.gen_dgrl`。download (bool or str, optional) – 如果資料集未找到,是否應該下載。預設為
True。也可以將 download 設定為"force",在這種情況下將覆蓋已下載的資料。sampler (Sampler, optional) – 要使用的取樣器。如果未提供,將使用預設的 RandomSampler()。
writer (Writer, optional) – 要使用的寫入器。如果未提供,將使用預設的 RoundRobinWriter()。
collate_fn (callable, optional) – 合併樣本列表以形成 Tensor(s)/輸出的小批次。在從對映式資料集批次載入時使用。
pin_memory (bool) – 是否應在 rb 樣本上呼叫 pin_memory()。
prefetch (int, optional) – 使用多執行緒預取下一個批次的數量。
transform (Transform, optional) – 在呼叫 sample() 時執行的轉換。要鏈式應用轉換,請使用
Compose類。
- 變數:
available_datasets – 可接受的待下載條目列表。這些名稱對應於 huggingface 資料集倉庫中的目錄路徑。如果可能,該列表將從 huggingface 動態檢索。如果沒有網際網路連線,則將使用快取的版本。
示例
>>> import torch >>> torch.manual_seed(0) >>> from torchrl.data.datasets import GenDGRLExperienceReplay >>> d = GenDGRLExperienceReplay("bigfish-1M_E", batch_size=32) >>> for batch in d: ... break >>> print(batch)
- add(data: TensorDictBase) int¶
向經驗回放緩衝區新增單個元素。
- 引數:
data (Any) – 要新增到經驗回放緩衝區的資料
- 返回:
資料在經驗回放緩衝區中的索引。
- append_transform(transform: Transform, *, invert: bool = False) ReplayBuffer¶
在末尾追加轉換。
呼叫 sample 時,轉換會按順序應用。
- 引數:
transform (Transform) – 要追加的轉換
- 關鍵字引數:
invert (bool, optional) – 如果為
True,則轉換將被反轉(forward 呼叫將在寫入時呼叫,inverse 呼叫將在讀取時呼叫)。預設為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¶
資料集的路徑,包括拆分。
- property data_path_root¶
資料集根目錄的路徑。
- delete()¶
從磁碟刪除資料集儲存。
- dumps(path)¶
將經驗回放緩衝區儲存到磁碟上的指定路徑。
- 引數:
path (Path or str) – 儲存經驗回放緩衝區的路徑。
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()¶
清空經驗回放緩衝區並將游標重置為 0。
- extend(tensordicts: TensorDictBase) Tensor¶
使用可迭代物件中包含的一個或多個元素擴充套件經驗回放緩衝區。
如果存在,將呼叫 inverse 轉換。`
- 引數:
data (iterable) – 要新增到經驗回放緩衝區的資料集合。
- 返回:
新增到經驗回放緩衝區的資料的索引。
警告
extend()在處理值列表時可能具有模糊的簽名,可以將其解釋為 PyTree(在這種情況下,列表中的所有元素都將放入儲存中 PyTree 的切片中)或要一次新增一個的值列表。為了解決這個問題,TorchRL 明確區分了 list 和 tuple:tuple 將被視為 PyTree,list(在根級別)將被解釋為堆疊中的值,用於一次向緩衝區新增一個。對於ListStorage例項,只能提供未繫結的元素(不能是 PyTrees)。
- insert_transform(index: int, transform: Transform, *, invert: bool = False) ReplayBuffer¶
插入轉換。
呼叫 sample 時,轉換會按順序執行。
- 引數:
index (int) – 插入轉換的位置。
transform (Transform) – 要追加的轉換
- 關鍵字引數:
invert (bool, optional) – 如果為
True,則轉換將被反轉(forward 呼叫將在寫入時呼叫,inverse 呼叫將在讀取時呼叫)。預設為False。
- loads(path)¶
從給定路徑載入經驗回放緩衝區狀態。
緩衝區應具有匹配的元件,並且是使用
dumps()儲存的。- 引數:
path (Path or str) – 經驗回放緩衝區儲存的路徑。
有關更多資訊,請參閱
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¶
預處理資料集並返回一個包含格式化資料的新儲存。
資料轉換必須是單元的(作用於資料集的單個樣本)。
引數和關鍵字引數會被轉發到
map()。隨後可以使用
delete()刪除資料集。- 關鍵字引數:
dest (path or equivalent) – 新資料集位置的路徑。
num_frames (int, optional) – 如果提供,則僅轉換前 num_frames 幀。這對於初步除錯轉換很有用。
返回: 一個新的儲存,可在
ReplayBuffer例項中使用。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])¶
為儲存註冊一個載入鉤子。
注意
當前在儲存經驗回放緩衝區時,鉤子不會被序列化:每次建立緩衝區時都必須手動重新初始化它們。
- register_save_hook(hook: Callable[[Any], Any])¶
為儲存註冊一個儲存鉤子。
注意
當前在儲存經驗回放緩衝區時,鉤子不會被序列化:每次建立緩衝區時都必須手動重新初始化它們。
- sample(batch_size: int | None = None, return_info: bool = False, include_info: bool =None) TensorDictBase¶
從經驗回放緩衝區取樣一個批次的資料。
使用 Sampler 取樣索引,並從 Storage 中檢索它們。
- 引數:
batch_size (int, optional) – 要收集的資料的大小。如果未提供,此方法將根據取樣器指示的批次大小進行取樣。
return_info (bool) – 是否返回 info。如果為 True,結果是一個 tuple (data, info)。如果為 False,結果是 data。
- 返回:
一個包含在經驗回放緩衝區中選定的批次資料的 tensordict。如果設定了 return_info 標誌為 True,則返回一個包含此 tensordict 和 info 的 tuple。
- set_storage(storage: Storage, collate_fn: Callable | None = None)¶
在回放緩衝區中設定新的儲存並返回先前的儲存。
- 引數:
storage (Storage) – 緩衝區的新的儲存。
collate_fn (callable, optional) – 如果提供,則 collate_fn 將設定為此值。否則它將被重置為預設值。
- property write_count¶
透過 add 和 extend 操作寫入緩衝區中的專案總數。