| # # Copyright 2024 Bytedance Ltd. and/or its affiliates | |
| # # | |
| # # Licensed under the Apache License, Version 2.0 (the "License"); | |
| # # you may not use this file except in compliance with the License. | |
| # # You may obtain a copy of the License at | |
| # # | |
| # # http://www.apache.org/licenses/LICENSE-2.0 | |
| # # | |
| # # Unless required by applicable law or agreed to in writing, software | |
| # # distributed under the License is distributed on an "AS IS" BASIS, | |
| # # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | |
| # # See the License for the specific language governing permissions and | |
| # # limitations under the License. | |
| # """ | |
| # Implement base data transfer protocol between any two functions, modules. | |
| # We can subclass Protocol to define more detailed batch info with specific keys | |
| # """ | |
| # import copy | |
| # import io | |
| # import os | |
| # import pickle | |
| # from collections import defaultdict | |
| # from dataclasses import dataclass, field | |
| # from typing import Any, Callable, Optional, Union | |
| # import numpy as np | |
| # import ray | |
| # import torch | |
| # from numpy.typing import NDArray | |
| # from tensordict import TensorDict | |
| # from torch.distributed import ProcessGroup | |
| # from torch.utils.data import DataLoader | |
| # from .utils.py_functional import union_two_dict | |
| # try: | |
| # import tensordict | |
| # tensordict.set_lazy_legacy(False).set() | |
| # except Exception: | |
| # pass | |
| # __all__ = ["DataProto", "union_tensor_dict"] | |
| # # ====== Feature switches (default OFF to keep original behavior) ====== | |
| # DP_SMART_UNION = 1 # per-sample smart union for non-tensors | |
| # DP_AUTO_ALIGN = 1 # auto align non_tensors length to batch_size | |
| # DP_RELAX_FROM_SINGLE = 1 # from_single_dict accepts non-ndarray by casting to object | |
| # def pad_dataproto_to_divisor(data: "DataProto", size_divisor: int) -> tuple["DataProto", int]: | |
| # """Pad a DataProto to size divisible by size_divisor""" | |
| # assert isinstance(data, DataProto), "data must be a DataProto" | |
| # if len(data) % size_divisor != 0: | |
| # pad_size = size_divisor - len(data) % size_divisor | |
| # padding_protos = [] | |
| # remaining_pad = pad_size | |
| # while remaining_pad > 0: | |
| # take_size = min(remaining_pad, len(data)) | |
| # padding_protos.append(data[:take_size]) | |
| # remaining_pad -= take_size | |
| # data_padded = DataProto.concat([data] + padding_protos) | |
| # else: | |
| # pad_size = 0 | |
| # data_padded = data | |
| # return data_padded, pad_size | |
| # def unpad_dataproto(data: "DataProto", pad_size: int) -> "DataProto": | |
| # if pad_size != 0: | |
| # data = data[:-pad_size] | |
| # return data | |
| # def union_tensor_dict(tensor_dict1: TensorDict, tensor_dict2: TensorDict) -> TensorDict: | |
| # """Union two tensordicts.""" | |
| # if tensor_dict1.batch_size != tensor_dict2.batch_size: | |
| # raise ValueError( | |
| # f"Two tensor dict must have identical batch size. Got {tensor_dict1.batch_size} and {tensor_dict2.batch_size}" | |
| # ) | |
| # for key in tensor_dict2.keys(): | |
| # if key in tensor_dict1 and not torch.equal(tensor_dict1[key], tensor_dict2[key]): | |
| # raise ValueError(f"Key already exists: {key}.") | |
| # tensor_dict1[key] = tensor_dict2[key] | |
| # return tensor_dict1 | |
| # # ---------------------- non-tensor union (two modes) ---------------------- | |
| # # Strict mode (original behavior) | |
| # def _union_numpy_dict_strict(tensor_dict1: dict[str, NDArray], tensor_dict2: dict[str, NDArray]) -> dict[str, NDArray]: | |
| # for key in tensor_dict2.keys(): | |
| # if key in tensor_dict1: | |
| # assert isinstance(tensor_dict2[key], np.ndarray) | |
| # assert isinstance(tensor_dict1[key], np.ndarray) | |
| # if not np.all(tensor_dict1[key] == tensor_dict2[key]): | |
| # raise ValueError(f"Key already exists: {key}.") | |
| # tensor_dict1[key] = tensor_dict2[key] | |
| # return tensor_dict1 | |
| # # Smart mode (opt-in): per-sample merge with None tolerance | |
| # def _to_obj_1d(x) -> NDArray: | |
| # if isinstance(x, np.ndarray) and x.dtype == np.dtype(object): | |
| # return x | |
| # return np.array(x, dtype=object) | |
| # def _elem_equal(x: Any, y: Any) -> bool: | |
| # if x is y: | |
| # return True | |
| # if isinstance(x, str) and isinstance(y, str): | |
| # return x == y | |
| # if isinstance(x, torch.Tensor) and isinstance(y, torch.Tensor): | |
| # try: | |
| # return x.shape == y.shape and torch.equal(x.cpu(), y.cpu()) | |
| # except Exception: | |
| # return False | |
| # if isinstance(x, np.ndarray) and isinstance(y, np.ndarray) and x.dtype != object and y.dtype != object: | |
| # try: | |
| # return x.shape == y.shape and np.array_equal(x, y) | |
| # except Exception: | |
| # return False | |
| # try: | |
| # return x == y | |
| # except Exception: | |
| # return False | |
| # def _union_numpy_dict_smart(tensor_dict1: dict[str, NDArray], tensor_dict2: dict[str, NDArray]) -> dict[str, NDArray]: | |
| # """ | |
| # Per-sample merge: | |
| # - missing side (None) -> take the other; | |
| # - both non-None and equal -> keep left; | |
| # - both non-None and not equal -> raise with conflict indices (first few). | |
| # """ | |
| # for key, v2 in tensor_dict2.items(): | |
| # if key not in tensor_dict1: | |
| # tensor_dict1[key] = v2 | |
| # continue | |
| # v1 = tensor_dict1[key] | |
| # a1 = _to_obj_1d(v1) | |
| # a2 = _to_obj_1d(v2) | |
| # if a1.shape != a2.shape: | |
| # raise ValueError(f"Key '{key}' has different shapes: {a1.shape} vs {a2.shape}") | |
| # merged = np.empty_like(a1, dtype=object) | |
| # conflict_idx = [] | |
| # for i in range(a1.shape[0]): | |
| # x, y = a1[i], a2[i] | |
| # if x is None and y is None: | |
| # merged[i] = None | |
| # elif x is None: | |
| # merged[i] = y | |
| # elif y is None: | |
| # merged[i] = x | |
| # else: | |
| # if _elem_equal(x, y): | |
| # merged[i] = x | |
| # else: | |
| # conflict_idx.append(i) | |
| # merged[i] = x # keep left, but we still error after loop | |
| # if conflict_idx: | |
| # head = conflict_idx[:5] | |
| # raise ValueError(f"Key already exists and conflicts on sample indices {head} (key='{key}').") | |
| # tensor_dict1[key] = merged | |
| # return tensor_dict1 | |
| # # Public entry: choose by switch (default strict/original) | |
| # def union_numpy_dict(tensor_dict1: dict[str, NDArray], tensor_dict2: dict[str, NDArray]) -> dict[str, NDArray]: | |
| # if DP_SMART_UNION: | |
| # return _union_numpy_dict_smart(tensor_dict1, tensor_dict2) | |
| # else: | |
| # return _union_numpy_dict_strict(tensor_dict1, tensor_dict2) | |
| # # ---------------------- helpers ---------------------- | |
| # def batch_collate(features: list[dict[str, Any]]) -> dict[str, list[Any]]: | |
| # if len(features) == 0: | |
| # return {} | |
| # batch_features = defaultdict(list) | |
| # for feature in features: | |
| # for key, value in feature.items(): | |
| # batch_features[key].append(value) | |
| # return batch_features | |
| # def fold_batch_dim(data: "DataProto", new_batch_size: int): | |
| # """ | |
| # Fold a batch dim from [bsz, xxx] into [new_bsz, bsz // new_bsz, xxx] | |
| # """ | |
| # batch_size = data.batch.batch_size[0] | |
| # assert batch_size % new_batch_size == 0 | |
| # tensor: TensorDict = data.batch | |
| # non_tensor = data.non_tensor_batch | |
| # tensor = tensor.view(new_batch_size, -1) | |
| # tensor.auto_batch_size_(batch_dims=1) | |
| # for key, value in non_tensor.items(): | |
| # non_tensor[key] = np.reshape(value, newshape=(new_batch_size, -1, *value.shape[1:])) | |
| # return DataProto(batch=tensor, non_tensor_batch=non_tensor, meta_info=data.meta_info) | |
| # def collate_fn(data_items: list["DataProtoItem"]): | |
| # batch = [] | |
| # non_tensor_batch = [] | |
| # for data in data_items: | |
| # batch.append(data.batch) | |
| # non_tensor_batch.append(data.non_tensor_batch) | |
| # batch = torch.stack(batch).contiguous() | |
| # non_tensor_batch = batch_collate(non_tensor_batch) | |
| # non_tensor_batch = {key: np.array(value, dtype=object) for key, value in non_tensor_batch.items()} | |
| # return DataProto(batch=batch, non_tensor_batch=non_tensor_batch) | |
| # @dataclass | |
| # class DataProtoItem: | |
| # batch: Optional[TensorDict] = None | |
| # non_tensor_batch: dict[str, NDArray] = field(default_factory=dict) | |
| # meta_info: dict[str, Any] = field(default_factory=dict) | |
| # @dataclass | |
| # class DataProto: | |
| # """ | |
| # A DataProto is a data structure that aims to provide a standard protocol for data exchange between functions. | |
| # It contains a batch (TensorDict) and a meta_info (Dict). | |
| # """ | |
| # batch: Optional[TensorDict] = None | |
| # non_tensor_batch: dict[str, NDArray] = field(default_factory=dict) | |
| # meta_info: dict[str, Any] = field(default_factory=dict) | |
| # @staticmethod | |
| # def _ensure_obj_array(v) -> NDArray: | |
| # if isinstance(v, np.ndarray) and v.dtype == np.dtype(object): | |
| # return v | |
| # return np.array(v, dtype=object) | |
| # @classmethod | |
| # def _align_non_tensors_len(cls, non_tensors: dict[str, NDArray], batch_size: int) -> None: | |
| # """Align non_tensors along dim0 to batch_size (pad with None or truncate).""" | |
| # for k, v in list(non_tensors.items()): | |
| # v = cls._ensure_obj_array(v) | |
| # cur = v.shape[0] | |
| # if cur < batch_size: | |
| # pad_n = batch_size - cur | |
| # pad = np.array([None] * pad_n, dtype=object) | |
| # non_tensors[k] = np.concatenate([v, pad], axis=0) | |
| # elif cur > batch_size: | |
| # non_tensors[k] = v[:batch_size] | |
| # else: | |
| # non_tensors[k] = v | |
| # def __post_init__(self): | |
| # # Only auto-align when explicitly enabled; otherwise keep original strictness. | |
| # if DP_AUTO_ALIGN and self.batch is not None and len(self.non_tensor_batch) != 0: | |
| # self._align_non_tensors_len(self.non_tensor_batch, self.batch.batch_size[0]) | |
| # self.check_consistency() # perform necessary checking | |
| # def __len__(self) -> int: | |
| # if self.batch is not None: | |
| # return self.batch.batch_size[0] | |
| # elif self.non_tensor_batch is not None and len(self.non_tensor_batch) > 0: | |
| # pivot_key = list(self.non_tensor_batch.keys())[0] | |
| # return self.non_tensor_batch[pivot_key].shape[0] | |
| # else: | |
| # return 0 | |
| # def __getitem__( | |
| # self, item: Union[int, slice, list[int], np.ndarray, torch.Tensor] | |
| # ) -> Union["DataProto", "DataProtoItem"]: | |
| # if isinstance(item, slice): | |
| # return self.slice_select(item.start, item.stop, item.step) | |
| # if isinstance(item, (list, np.ndarray, torch.Tensor)): | |
| # return self.index_select(item) | |
| # if isinstance(item, (int, np.integer)): | |
| # tensor_data = self.batch[item] if self.batch is not None else None | |
| # non_tensor_data = {key: value[item] for key, value in self.non_tensor_batch.items()} | |
| # return DataProtoItem(batch=tensor_data, non_tensor_batch=non_tensor_data, meta_info=self.meta_info) | |
| # raise TypeError(f"Indexing with {type(item)} is not supported.") | |
| # def __getstate__(self) -> tuple[bytes, dict[str, NDArray], dict[str, Any]]: | |
| # if self.batch is not None: | |
| # batch_to_save: TensorDict = self.batch.contiguous() | |
| # batch_to_save: TensorDict = batch_to_save.consolidate() | |
| # else: | |
| # batch_to_save = None | |
| # buffer = io.BytesIO() | |
| # torch.save(batch_to_save, buffer) | |
| # buffer_bytes = buffer.getvalue() | |
| # return buffer_bytes, self.non_tensor_batch, self.meta_info | |
| # def __setstate__(self, data: tuple[bytes, dict[str, NDArray], dict[str, Any]]) -> None: | |
| # batch_deserialized_bytes, non_tensor_batch, meta_info = data | |
| # batch_deserialized = io.BytesIO(batch_deserialized_bytes) | |
| # batch = torch.load(batch_deserialized, weights_only=False, map_location="cpu") | |
| # self.batch = batch | |
| # self.non_tensor_batch = non_tensor_batch | |
| # self.meta_info = meta_info | |
| # def save_to_disk(self, filepath: str) -> None: | |
| # with open(filepath, "wb") as f: | |
| # pickle.dump(self, f) | |
| # @staticmethod | |
| # def load_from_disk(filepath: str) -> "DataProto": | |
| # with open(filepath, "rb") as f: | |
| # data = pickle.load(f) | |
| # return data | |
| # def print_size(self, prefix: str = "") -> None: | |
| # size_of_tensordict = 0 | |
| # if self.batch is not None: | |
| # for tensor in self.batch.values(): | |
| # if isinstance(tensor, torch.Tensor): | |
| # size_of_tensordict += tensor.element_size() * tensor.numel() | |
| # size_of_numpy_array = 0 | |
| # for value in self.non_tensor_batch.values(): | |
| # size_of_numpy_array += value.nbytes | |
| # size_of_numpy_array /= 1024**3 | |
| # size_of_tensordict /= 1024**3 | |
| # message = f"Size of tensordict: {size_of_tensordict} GB, size of non_tensor_batch: {size_of_numpy_array} GB." | |
| # print({prefix}, {message}) | |
| # def check_consistency(self): | |
| # """Check the consistency of the DataProto. Mainly for batch and non_tensor_batch""" | |
| # if self.batch is not None: | |
| # assert len(self.batch.batch_size) == 1, "only support num_batch_dims=1" | |
| # if self.batch is not None and len(self.non_tensor_batch) != 0: | |
| # assert len(self.batch.batch_size) == 1, "only support num_batch_dims=1 when non_tensor_batch is not empty." | |
| # batch_size = self.batch.batch_size[0] | |
| # # Only auto-align when enabled; otherwise keep original assertion behavior | |
| # if DP_AUTO_ALIGN: | |
| # self._align_non_tensors_len(self.non_tensor_batch, batch_size) | |
| # for key, value in self.non_tensor_batch.items(): | |
| # assert len(value) == batch_size, f"key {key} length {len(value)} is not equal to bsz {batch_size}." | |
| # @classmethod | |
| # def from_single_dict( | |
| # cls, | |
| # data: dict[str, Union[torch.Tensor, NDArray]], | |
| # meta_info: Optional[dict[str, Any]] = None, | |
| # ) -> "DataProto": | |
| # tensors, non_tensors = {}, {} | |
| # for key, value in data.items(): | |
| # if isinstance(value, torch.Tensor): | |
| # tensors[key] = value | |
| # elif isinstance(value, np.ndarray): | |
| # non_tensors[key] = value | |
| # else: | |
| # if DP_RELAX_FROM_SINGLE: | |
| # non_tensors[key] = np.array(value, dtype=object) | |
| # else: | |
| # raise ValueError(f"Unsupported type in data {type(value)}") | |
| # return DataProto.from_dict(tensors=tensors, non_tensors=non_tensors, meta_info=meta_info) | |
| # @classmethod | |
| # def from_dict( | |
| # cls, | |
| # tensors: Optional[dict[str, torch.Tensor]] = None, | |
| # non_tensors: Optional[dict[str, NDArray]] = None, | |
| # meta_info: Optional[dict[str, Any]] = None, | |
| # num_batch_dims: int = 1, | |
| # ) -> "DataProto": | |
| # """Create a DataProto from a dict of tensors.""" | |
| # assert num_batch_dims > 0, "num_batch_dims must be greater than zero" | |
| # if non_tensors is not None: | |
| # assert num_batch_dims == 1, "only support num_batch_dims=1 when non_tensors is not None." | |
| # tensors = tensors or {} | |
| # non_tensors = non_tensors or {} | |
| # meta_info = meta_info or {} | |
| # assert isinstance(tensors, dict) and isinstance(non_tensors, dict) and isinstance(meta_info, dict) | |
| # # get and check batch size | |
| # batch_size = None | |
| # pivot_key = None | |
| # for key, tensor in tensors.items(): | |
| # if batch_size is None: | |
| # batch_size = tensor.shape[:num_batch_dims] | |
| # pivot_key = key | |
| # else: | |
| # current_batch = tensor.shape[:num_batch_dims] | |
| # assert batch_size == current_batch, ( | |
| # f"Not all the tensor in tensors have the same batch size with batch_dims={num_batch_dims}. " | |
| # f"Got {pivot_key} has {batch_size}, {key} has {current_batch}." | |
| # ) | |
| # # keep original: normalize non_tensors to object ndarray for consistency | |
| # for key, value in non_tensors.items(): | |
| # if not isinstance(value, np.ndarray) or value.dtype != np.dtype(object): | |
| # non_tensors[key] = np.array(value, dtype=object) | |
| # tensor_dict = TensorDict(source=tensors, batch_size=batch_size) if tensors else None | |
| # return cls(batch=tensor_dict, non_tensor_batch=non_tensors, meta_info=meta_info) | |
| # def to(self, device: torch.device, non_blocking: bool = False) -> "DataProto": | |
| # """Move the batch to device""" | |
| # if self.batch is not None: | |
| # self.batch = self.batch.to(device, non_blocking=non_blocking) | |
| # return self | |
| # def select( | |
| # self, | |
| # batch_keys: Optional[list[str]] = None, | |
| # non_tensor_batch_keys: Optional[list[str]] = None, | |
| # meta_info_keys: Optional[list[str]] = None, | |
| # deepcopy: bool = False, | |
| # ) -> "DataProto": | |
| # """Select a subset of the DataProto via batch_keys and meta_info_keys""" | |
| # if batch_keys is not None: | |
| # batch_keys = tuple(filter(lambda k: k in self.batch, batch_keys)) | |
| # sub_batch = self.batch.select(*batch_keys) | |
| # else: | |
| # sub_batch = self.batch | |
| # if non_tensor_batch_keys is not None: | |
| # non_tensor_batch_keys = tuple(filter(lambda k: k in self.non_tensor_batch, non_tensor_batch_keys)) | |
| # non_tensor_batch = {k: v for k, v in self.non_tensor_batch.items() if k in non_tensor_batch_keys} | |
| # else: | |
| # non_tensor_batch = self.non_tensor_batch | |
| # if deepcopy: | |
| # non_tensor_batch = copy.deepcopy(non_tensor_batch) | |
| # if meta_info_keys is not None: | |
| # meta_info_keys = tuple(filter(lambda k: k in self.meta_info, meta_info_keys)) | |
| # sub_meta_info = {k: v for k, v in self.meta_info.items() if k in meta_info_keys} | |
| # else: | |
| # sub_meta_info = self.meta_info | |
| # if deepcopy: | |
| # sub_meta_info = copy.deepcopy(sub_meta_info) | |
| # return DataProto(batch=sub_batch, non_tensor_batch=non_tensor_batch, meta_info=sub_meta_info) | |
| # def index_select(self, index: Union[list[int], NDArray, torch.Tensor]) -> "DataProto": | |
| # """Select a subset via index.""" | |
| # if isinstance(index, list): | |
| # index = np.array(index, dtype=bool if isinstance(index[0], bool) else np.int32) | |
| # elif isinstance(index, torch.Tensor): | |
| # index = index.detach().cpu().numpy() | |
| # tensor_data = self.batch[index] if self.batch is not None else None | |
| # non_tensor_data = {key: value[index] for key, value in self.non_tensor_batch.items()} | |
| # return DataProto(batch=tensor_data, non_tensor_batch=non_tensor_data, meta_info=self.meta_info) | |
| # def slice_select( | |
| # self, start: Optional[int] = None, end: Optional[int] = None, step: Optional[int] = None | |
| # ) -> "DataProto": | |
| # """Select a subset via slice.""" | |
| # index = slice(start, end, step) | |
| # tensor_data = self.batch[index] if self.batch is not None else None | |
| # non_tensor_data = {key: value[index] for key, value in self.non_tensor_batch.items()} | |
| # return DataProto(batch=tensor_data, non_tensor_batch=non_tensor_data, meta_info=self.meta_info) | |
| # def pop( | |
| # self, | |
| # batch_keys: Optional[list[str]] = None, | |
| # non_tensor_batch_keys: Optional[list[str]] = None, | |
| # meta_info_keys: Optional[list[str]] = None, | |
| # ) -> "DataProto": | |
| # """Pop a subset of the DataProto via keys""" | |
| # assert batch_keys is not None | |
| # non_tensor_batch_keys = non_tensor_batch_keys or [] | |
| # meta_info_keys = meta_info_keys or [] | |
| # tensors = {} | |
| # for key in filter(lambda k: k in self.batch, batch_keys): | |
| # tensors[key] = self.batch.pop(key) | |
| # non_tensors = {} | |
| # for key in filter(lambda k: k in self.non_tensor_batch, non_tensor_batch_keys): | |
| # non_tensors[key] = self.non_tensor_batch.pop(key) | |
| # meta_info = {} | |
| # for key in filter(lambda k: k in self.meta_info, meta_info_keys): | |
| # meta_info[key] = self.meta_info.pop(key) | |
| # return DataProto.from_dict(tensors=tensors, non_tensors=non_tensors, meta_info=meta_info) | |
| # def rename( | |
| # self, old_keys: Optional[Union[str, list[str]]] = None, new_keys: Optional[Union[str, list[str]]] = None | |
| # ) -> "DataProto": | |
| # """Rename keys in the batch only.""" | |
| # def validate_input(keys): | |
| # if keys is not None: | |
| # if isinstance(keys, str): | |
| # keys = [keys] | |
| # elif isinstance(keys, list): | |
| # pass | |
| # else: | |
| # raise TypeError(f"keys must be a list or a string, but got {type(keys)}") | |
| # return keys | |
| # old_keys = validate_input(old_keys) | |
| # new_keys = validate_input(new_keys) | |
| # if len(new_keys) != len(old_keys): | |
| # raise ValueError( | |
| # f"new_keys and old_keys must have the same length, but got {len(new_keys)} and {len(old_keys)}" | |
| # ) | |
| # self.batch.rename_key_(tuple(old_keys), tuple(new_keys)) | |
| # return self | |
| # def union(self, other: "DataProto") -> "DataProto": | |
| # """Union with another DataProto. Union batch and meta_info separately.""" | |
| # self.batch = union_tensor_dict(self.batch, other.batch) | |
| # self.non_tensor_batch = union_numpy_dict(self.non_tensor_batch, other.non_tensor_batch) | |
| # self.meta_info = union_two_dict(self.meta_info, other.meta_info) | |
| # return self | |
| # def make_iterator( | |
| # self, mini_batch_size: int, epochs: int, seed: int = None, dataloader_kwargs: dict[str, Any] = None | |
| # ): | |
| # """Make an iterator from the DataProto.""" | |
| # assert self.batch.batch_size[0] % mini_batch_size == 0, f"{self.batch.batch_size[0]} % {mini_batch_size} != 0" | |
| # if seed is not None: | |
| # generator = torch.Generator() | |
| # generator.manual_seed(seed) | |
| # else: | |
| # generator = None | |
| # dataloader_kwargs = dataloader_kwargs or {} | |
| # assert isinstance(dataloader_kwargs, dict) | |
| # train_dataloader = DataLoader( | |
| # dataset=self, | |
| # batch_size=mini_batch_size, | |
| # collate_fn=collate_fn, | |
| # generator=generator, | |
| # **dataloader_kwargs, | |
| # ) | |
| # def get_data(): | |
| # for _ in range(epochs): | |
| # for data in train_dataloader: | |
| # setattr(data, "meta_info", self.meta_info) | |
| # yield data | |
| # return iter(get_data()) | |
| # def chunk(self, chunks: int) -> list["DataProto"]: | |
| # """Split the batch among dim=0 into chunks.""" | |
| # assert len(self) % chunks == 0, ( | |
| # f"only support equal chunk. Got size of DataProto {len(self)} and chunk {chunks}." | |
| # ) | |
| # if self.batch is not None: | |
| # batch_lst = self.batch.chunk(chunks=chunks, dim=0) | |
| # else: | |
| # batch_lst = [None for _ in range(chunks)] | |
| # non_tensor_batch_lst = [{} for _ in range(chunks)] | |
| # for key, value in self.non_tensor_batch.items(): | |
| # non_tensor_lst = np.array_split(value, chunks) | |
| # for i in range(chunks): | |
| # non_tensor_batch_lst[i][key] = non_tensor_lst[i] | |
| # return [ | |
| # DataProto(batch=batch_lst[i], non_tensor_batch=non_tensor_batch_lst[i], meta_info=self.meta_info) | |
| # for i in range(chunks) | |
| # ] | |
| # def split(self, split_size: int) -> list["DataProto"]: | |
| # """Split the batch among dim=0 into chunks.""" | |
| # assert len(self) % split_size == 0, ( | |
| # f"only support equal split. Got size of DataProto {len(self)} and split {split_size}." | |
| # ) | |
| # chunks = len(self) // split_size | |
| # return self.chunk(chunks) | |
| # @staticmethod | |
| # def concat(data: list["DataProto"]) -> "DataProto": | |
| # """Concat a list of DataProto.""" | |
| # batch_lst = [batch.batch for batch in data] | |
| # new_batch = torch.cat(batch_lst, dim=0) if batch_lst[0] is not None else None | |
| # non_tensor_batch = batch_collate([d.non_tensor_batch for d in data]) | |
| # for key, value in non_tensor_batch.items(): | |
| # non_tensor_batch[key] = np.concatenate(value, axis=0) | |
| # return DataProto(batch=new_batch, non_tensor_batch=non_tensor_batch, meta_info=data[0].meta_info) | |
| # def reorder(self, indices: torch.Tensor) -> None: | |
| # """In-place reorder by indices.""" | |
| # indices_np = indices.detach().numpy() | |
| # self.batch = self.batch[indices] | |
| # self.non_tensor_batch = {key: value[indices_np] for key, value in self.non_tensor_batch.items()} | |
| # def repeat(self, repeat_times: int, interleave: bool = True) -> "DataProto": | |
| # """Repeat the batch data a specified number of times.""" | |
| # if self.batch is not None: | |
| # if interleave: | |
| # repeated_tensors = { | |
| # key: tensor.repeat_interleave(repeat_times, dim=0) for key, tensor in self.batch.items() | |
| # } | |
| # else: | |
| # repeated_tensors = { | |
| # key: tensor.unsqueeze(0).expand(repeat_times, *tensor.shape).reshape(-1, *tensor.shape[1:]) | |
| # for key, tensor in self.batch.items() | |
| # } | |
| # repeated_batch = TensorDict( | |
| # source=repeated_tensors, | |
| # batch_size=(self.batch.batch_size[0] * repeat_times,), | |
| # ) | |
| # else: | |
| # repeated_batch = None | |
| # repeated_non_tensor_batch = {} | |
| # for key, value in self.non_tensor_batch.items(): | |
| # if interleave: | |
| # repeated_non_tensor_batch[key] = np.repeat(value, repeat_times, axis=0) | |
| # else: | |
| # repeated_non_tensor_batch[key] = np.tile(value, (repeat_times,) + (1,) * (value.ndim - 1)) | |
| # return DataProto( | |
| # batch=repeated_batch, | |
| # non_tensor_batch=repeated_non_tensor_batch, | |
| # meta_info=self.meta_info, | |
| # ) | |
| # @dataclass | |
| # class DataProtoFuture: | |
| # """ | |
| # DataProtoFuture aims to eliminate actual data fetching on driver. | |
| # """ | |
| # collect_fn: Callable | |
| # futures: list[ray.ObjectRef] | |
| # dispatch_fn: Callable = None | |
| # @staticmethod | |
| # def concat(data: list[ray.ObjectRef]) -> "DataProtoFuture": | |
| # output = DataProtoFuture(collect_fn=DataProto.concat, futures=data) | |
| # return output | |
| # def chunk(self, chunks: int) -> list["DataProtoFuture"]: | |
| # from functools import partial | |
| # arg_future_lst = [] | |
| # for i in range(chunks): | |
| # # note that we can't directly pass i and chunks | |
| # def dispatch_fn(x, i, chunks): | |
| # return x.chunk(chunks=chunks)[i] | |
| # arg_future = DataProtoFuture( | |
| # collect_fn=self.collect_fn, dispatch_fn=partial(dispatch_fn, i=i, chunks=chunks), futures=self.futures | |
| # ) | |
| # arg_future_lst.append(arg_future) | |
| # return arg_future_lst | |
| # def get(self): | |
| # outputs = ray.get(self.futures) # dp_size | |
| # for output in outputs: | |
| # assert isinstance(output, DataProto) | |
| # outputs = self.collect_fn(outputs) # select dp, concat | |
| # if self.dispatch_fn is not None: | |
| # outputs = self.dispatch_fn(outputs) # split in batch dim, select using dp | |
| # return outputs | |
| # def allgather_dict_tensors( | |
| # tensors: Union[dict[str, torch.Tensor], TensorDict], size: int, group: ProcessGroup, dim: int = 0 | |
| # ) -> Union[dict[str, torch.Tensor], TensorDict]: | |
| # """ | |
| # TODO: optimize this. | |
| # - We can use async ops | |
| # - We can use only one allgather | |
| # """ | |
| # if isinstance(tensors, TensorDict): | |
| # is_tensor_dict = True | |
| # tensors_as_dict = tensors.to_dict() | |
| # else: | |
| # tensors_as_dict = tensors | |
| # is_tensor_dict = False | |
| # output = {} | |
| # sorted_keys = sorted(tensors_as_dict.keys()) | |
| # for key in sorted_keys: | |
| # value = tensors_as_dict[key] | |
| # output[key] = [torch.empty_like(value) for _ in range(size)] | |
| # torch.distributed.all_gather(output[key], value, group=group, async_op=False) | |
| # output[key] = torch.cat(output[key], dim=dim) | |
| # if is_tensor_dict: | |
| # output = TensorDict(source=output, batch_size=tensors.batch_size[0] * size) | |
| # return output | |
| # def all_gather_data_proto(data: DataProto, size: int, group: ProcessGroup) -> None: | |
| # # Note that this is an inplace operator just like torch.distributed.all_gather | |
| # prev_device = data.batch.device | |
| # data.batch = data.batch.cuda(device=torch.cuda.current_device()) | |
| # data.batch = allgather_dict_tensors(data.batch.contiguous(), size=size, group=group, dim=0) | |
| # data.batch = data.batch.to(prev_device) | |
| # # all gather non_tensor_batch | |
| # all_non_tensor_batch = [None for _ in range(size)] | |
| # torch.distributed.all_gather_object(all_non_tensor_batch, data.non_tensor_batch, group=group) | |
| # data.non_tensor_batch = {k: np.concatenate([d[k] for d in all_non_tensor_batch]) for k in data.non_tensor_batch} | |
| # Copyright 2024 Bytedance Ltd. and/or its affiliates | |
| # | |
| # Licensed under the Apache License, Version 2.0 (the "License"); | |
| # you may not use this file except in compliance with the License. | |
| # You may obtain a copy of the License at | |
| # | |
| # http://www.apache.org/licenses/LICENSE-2.0 | |
| # | |
| # Unless required by applicable law or agreed to in writing, software | |
| # distributed under the License is distributed on an "AS IS" BASIS, | |
| # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | |
| # See the License for the specific language governing permissions and | |
| # limitations under the License. | |
| """ | |
| Implement base data transfer protocol between any two functions, modules. | |
| We can subclass Protocol to define more detailed batch info with specific keys | |
| """ | |
| import copy | |
| import io | |
| import pickle | |
| from collections import defaultdict | |
| from dataclasses import dataclass, field | |
| from typing import Any, Callable, Optional, Union | |
| import numpy as np | |
| import ray | |
| import torch | |
| from numpy.typing import NDArray | |
| from tensordict import TensorDict | |
| from torch.distributed import ProcessGroup | |
| from torch.utils.data import DataLoader | |
| from .utils.py_functional import union_two_dict | |
| try: | |
| import tensordict | |
| tensordict.set_lazy_legacy(False).set() | |
| except Exception: | |
| pass | |
| __all__ = ["DataProto", "union_tensor_dict"] | |
| def pad_dataproto_to_divisor(data: "DataProto", size_divisor: int) -> tuple["DataProto", int]: | |
| """Pad a DataProto to size divisible by size_divisor | |
| Args: | |
| data (DataProto): the unpadded DataProto | |
| size_divisor (int): size divisor | |
| Returns: | |
| data (DataProto): the padded DataProto | |
| pad_size (int) | |
| """ | |
| assert isinstance(data, DataProto), "data must be a DataProto" | |
| if len(data) % size_divisor != 0: | |
| pad_size = size_divisor - len(data) % size_divisor | |
| padding_protos = [] | |
| remaining_pad = pad_size | |
| while remaining_pad > 0: | |
| take_size = min(remaining_pad, len(data)) | |
| padding_protos.append(data[:take_size]) | |
| remaining_pad -= take_size | |
| data_padded = DataProto.concat([data] + padding_protos) | |
| else: | |
| pad_size = 0 | |
| data_padded = data | |
| return data_padded, pad_size | |
| def unpad_dataproto(data: "DataProto", pad_size: int) -> "DataProto": | |
| if pad_size != 0: | |
| data = data[:-pad_size] | |
| return data | |
| def union_tensor_dict(tensor_dict1: TensorDict, tensor_dict2: TensorDict) -> TensorDict: | |
| """Union two tensordicts.""" | |
| if tensor_dict1.batch_size != tensor_dict2.batch_size: | |
| raise ValueError( | |
| f"Two tensor dict must have identical batch size. Got {tensor_dict1.batch_size} and {tensor_dict2.batch_size}" | |
| ) | |
| for key in tensor_dict2.keys(): | |
| if key in tensor_dict1 and not torch.equal(tensor_dict1[key], tensor_dict2[key]): | |
| raise ValueError(f"Key already exists: {key}.") | |
| tensor_dict1[key] = tensor_dict2[key] | |
| return tensor_dict1 | |
| def union_numpy_dict(tensor_dict1: dict[str, NDArray], tensor_dict2: dict[str, NDArray]) -> dict[str, NDArray]: | |
| for key in tensor_dict2.keys(): | |
| if key in tensor_dict1: | |
| assert isinstance(tensor_dict2[key], np.ndarray) | |
| assert isinstance(tensor_dict1[key], np.ndarray) | |
| if not np.all(tensor_dict1[key] == tensor_dict2[key]): | |
| raise ValueError(f"Key already exists: {key}.") | |
| tensor_dict1[key] = tensor_dict2[key] | |
| return tensor_dict1 | |
| def batch_collate(features: list[dict[str, Any]]) -> dict[str, list[Any]]: | |
| if len(features) == 0: | |
| return {} | |
| batch_features = defaultdict(list) | |
| for feature in features: | |
| for key, value in feature.items(): | |
| batch_features[key].append(value) | |
| return batch_features | |
| def fold_batch_dim(data: "DataProto", new_batch_size: int): | |
| """ | |
| Fold a batch dim from [bsz, xxx] into [new_bsz, bsz // new_bsz, xxx] | |
| """ | |
| batch_size = data.batch.batch_size[0] | |
| assert batch_size % new_batch_size == 0 | |
| tensor: TensorDict = data.batch | |
| non_tensor = data.non_tensor_batch | |
| tensor = tensor.view(new_batch_size, -1) | |
| tensor.auto_batch_size_(batch_dims=1) | |
| for key, value in non_tensor.items(): | |
| non_tensor[key] = np.reshape(value, newshape=(new_batch_size, -1, *value.shape[1:])) | |
| return DataProto(batch=tensor, non_tensor_batch=non_tensor, meta_info=data.meta_info) | |
| def collate_fn(data_items: list["DataProtoItem"]): | |
| batch = [] | |
| non_tensor_batch = [] | |
| for data in data_items: | |
| batch.append(data.batch) | |
| non_tensor_batch.append(data.non_tensor_batch) | |
| batch = torch.stack(batch).contiguous() | |
| non_tensor_batch = batch_collate(non_tensor_batch) | |
| non_tensor_batch = {key: np.array(value, dtype=object) for key, value in non_tensor_batch.items()} | |
| return DataProto(batch=batch, non_tensor_batch=non_tensor_batch) | |
| class DataProtoItem: | |
| batch: Optional[TensorDict] = None | |
| non_tensor_batch: dict[str, NDArray] = field(default_factory=dict) | |
| meta_info: dict[str, Any] = field(default_factory=dict) | |
| class DataProto: | |
| """ | |
| A DataProto is a data structure that aims to provide a standard protocol for data exchange between functions. | |
| It contains a batch (TensorDict) and a meta_info (Dict). The batch is a TensorDict https://pytorch.org/tensordict/. | |
| TensorDict allows you to manipulate a dictionary of Tensors like a single Tensor. Ideally, the tensors with the | |
| same batch size should be put inside batch. | |
| """ | |
| batch: Optional[TensorDict] = None | |
| non_tensor_batch: dict[str, NDArray] = field(default_factory=dict) | |
| meta_info: dict[str, Any] = field(default_factory=dict) | |
| def __post_init__(self): | |
| self.check_consistency() # perform necessary checking | |
| def __len__(self) -> int: | |
| if self.batch is not None: | |
| return self.batch.batch_size[0] | |
| elif self.non_tensor_batch is not None and len(self.non_tensor_batch) > 0: | |
| pivot_key = list(self.non_tensor_batch.keys())[0] | |
| return self.non_tensor_batch[pivot_key].shape[0] | |
| else: | |
| return 0 | |
| def __getitem__( | |
| self, item: Union[int, slice, list[int], np.ndarray, torch.Tensor] | |
| ) -> Union["DataProto", "DataProtoItem"]: | |
| if isinstance(item, slice): | |
| return self.slice_select(item.start, item.stop, item.step) | |
| if isinstance(item, (list, np.ndarray, torch.Tensor)): | |
| return self.index_select(item) | |
| if isinstance(item, (int, np.integer)): | |
| tensor_data = self.batch[item] if self.batch is not None else None | |
| non_tensor_data = {key: value[item] for key, value in self.non_tensor_batch.items()} | |
| return DataProtoItem(batch=tensor_data, non_tensor_batch=non_tensor_data, meta_info=self.meta_info) | |
| raise TypeError(f"Indexing with {type(item)} is not supported.") | |
| def __getstate__(self) -> tuple[bytes, dict[str, NDArray], dict[str, Any]]: | |
| if self.batch is not None: | |
| batch_to_save: TensorDict = self.batch.contiguous() | |
| batch_to_save: TensorDict = batch_to_save.consolidate() | |
| else: | |
| batch_to_save = None | |
| buffer = io.BytesIO() | |
| torch.save(batch_to_save, buffer) | |
| buffer_bytes = buffer.getvalue() | |
| return buffer_bytes, self.non_tensor_batch, self.meta_info | |
| def __setstate__(self, data: tuple[bytes, dict[str, NDArray], dict[str, Any]]) -> None: | |
| batch_deserialized_bytes, non_tensor_batch, meta_info = data | |
| batch_deserialized = io.BytesIO(batch_deserialized_bytes) | |
| batch = torch.load(batch_deserialized, weights_only=False, map_location="cpu") | |
| self.batch = batch | |
| self.non_tensor_batch = non_tensor_batch | |
| self.meta_info = meta_info | |
| def save_to_disk(self, filepath: str) -> None: | |
| with open(filepath, "wb") as f: | |
| pickle.dump(self, f) | |
| def load_from_disk(filepath: str) -> "DataProto": | |
| with open(filepath, "rb") as f: | |
| data = pickle.load(f) | |
| return data | |
| def print_size(self, prefix: str = "") -> None: | |
| size_of_tensordict = 0 | |
| if self.batch is not None: | |
| for tensor in self.batch.values(): | |
| if isinstance(tensor, torch.Tensor): | |
| size_of_tensordict += tensor.element_size() * tensor.numel() | |
| size_of_numpy_array = 0 | |
| for value in self.non_tensor_batch.values(): | |
| size_of_numpy_array += value.nbytes | |
| size_of_numpy_array /= 1024**3 | |
| size_of_tensordict /= 1024**3 | |
| message = f"Size of tensordict: {size_of_tensordict} GB, size of non_tensor_batch: {size_of_numpy_array} GB." | |
| print({prefix}, {message}) | |
| def check_consistency(self): | |
| """Check the consistency of the DataProto. Mainly for batch and non_tensor_batch | |
| We expose this function as a public one so that user can call themselves directly | |
| """ | |
| if self.batch is not None: | |
| assert len(self.batch.batch_size) == 1, "only support num_batch_dims=1" | |
| if self.batch is not None and len(self.non_tensor_batch) != 0: | |
| # TODO: we can actually lift this restriction if needed | |
| assert len(self.batch.batch_size) == 1, "only support num_batch_dims=1 when non_tensor_batch is not empty." | |
| batch_size = self.batch.batch_size[0] | |
| #print(self.non_tensor_batch.items()) | |
| for key, value in self.non_tensor_batch.items(): | |
| assert len(value) == batch_size, f"key {key} length {len(value)} is not equal to bsz {batch_size}." | |
| def from_single_dict( | |
| cls, | |
| data: dict[str, Union[torch.Tensor, NDArray]], | |
| meta_info: Optional[dict[str, Any]] = None, | |
| ) -> "DataProto": | |
| tensors, non_tensors = {}, {} | |
| for key, value in data.items(): | |
| if isinstance(value, torch.Tensor): | |
| tensors[key] = value | |
| elif isinstance(value, np.ndarray): | |
| non_tensors[key] = value | |
| else: | |
| raise ValueError(f"Unsupported type in data {type(value)}") | |
| return DataProto.from_dict(tensors=tensors, non_tensors=non_tensors, meta_info=meta_info) | |
| def from_dict( | |
| cls, | |
| tensors: Optional[dict[str, torch.Tensor]] = None, | |
| non_tensors: Optional[dict[str, NDArray]] = None, | |
| meta_info: Optional[dict[str, Any]] = None, | |
| num_batch_dims: int = 1, | |
| ) -> "DataProto": | |
| """Create a DataProto from a dict of tensors. This assumes that | |
| 1. All the tensor in tensors have the same dim0 | |
| 2. Only dim0 is the batch dim | |
| """ | |
| assert num_batch_dims > 0, "num_batch_dims must be greater than zero" | |
| if non_tensors is not None: | |
| assert num_batch_dims == 1, "only support num_batch_dims=1 when non_tensors is not None." | |
| tensors = tensors or {} | |
| non_tensors = non_tensors or {} | |
| meta_info = meta_info or {} | |
| assert isinstance(tensors, dict) and isinstance(non_tensors, dict) and isinstance(meta_info, dict) | |
| # get and check batch size | |
| batch_size = None | |
| pivot_key = None | |
| for key, tensor in tensors.items(): | |
| if batch_size is None: | |
| batch_size = tensor.shape[:num_batch_dims] | |
| pivot_key = key | |
| else: | |
| current_batch = tensor.shape[:num_batch_dims] | |
| assert batch_size == current_batch, ( | |
| f"Not all the tensor in tensors have the same batch size with batch_dims={num_batch_dims}. " | |
| f"Got {pivot_key} has {batch_size}, {key} has {current_batch}." | |
| ) | |
| for key, value in non_tensors.items(): | |
| if not isinstance(value, np.ndarray) or value.dtype != np.dtype(object): | |
| non_tensors[key] = np.array(value, dtype=object) | |
| tensor_dict = TensorDict(source=tensors, batch_size=batch_size) if tensors else None | |
| return cls(batch=tensor_dict, non_tensor_batch=non_tensors, meta_info=meta_info) | |
| def to(self, device: torch.device, non_blocking: bool = False) -> "DataProto": | |
| """Move the batch to device | |
| Args: | |
| device (torch.device): the device to move to. | |
| non_blocking (bool, optional): whether to use non-blocking mode. Defaults to False. | |
| Returns: | |
| DataProto: the current DataProto. | |
| NOTE: remember to use torch.cuda.synchronize() after self.to("cpu") to avoid weird number | |
| """ | |
| if self.batch is not None: | |
| self.batch = self.batch.to(device, non_blocking=non_blocking) | |
| return self | |
| def select( | |
| self, | |
| batch_keys: Optional[list[str]] = None, | |
| non_tensor_batch_keys: Optional[list[str]] = None, | |
| meta_info_keys: Optional[list[str]] = None, | |
| deepcopy: bool = False, | |
| ) -> "DataProto": | |
| """Select a subset of the DataProto via batch_keys and meta_info_keys | |
| Args: | |
| batch_keys (list, optional): a list of strings indicating the keys in batch to select | |
| meta_info_keys (list, optional): a list of keys indicating the meta info to select | |
| Returns: | |
| DataProto: the DataProto with the selected batch_keys and meta_info_keys | |
| """ | |
| # TODO (zhangchi.usc1992) whether to copy | |
| if batch_keys is not None: | |
| batch_keys = tuple(filter(lambda k: k in self.batch, batch_keys)) | |
| sub_batch = self.batch.select(*batch_keys) | |
| else: | |
| sub_batch = self.batch | |
| if non_tensor_batch_keys is not None: | |
| # we must convert it to tuple to avoid the missing elements | |
| non_tensor_batch_keys = tuple(filter(lambda k: k in self.non_tensor_batch, non_tensor_batch_keys)) | |
| non_tensor_batch = {k: v for k, v in self.non_tensor_batch.items() if k in non_tensor_batch_keys} | |
| else: | |
| non_tensor_batch = self.non_tensor_batch | |
| if deepcopy: | |
| non_tensor_batch = copy.deepcopy(non_tensor_batch) | |
| if meta_info_keys is not None: | |
| meta_info_keys = tuple(filter(lambda k: k in self.meta_info, meta_info_keys)) | |
| sub_meta_info = {k: v for k, v in self.meta_info.items() if k in meta_info_keys} | |
| else: | |
| sub_meta_info = self.meta_info | |
| if deepcopy: | |
| sub_meta_info = copy.deepcopy(sub_meta_info) | |
| return DataProto(batch=sub_batch, non_tensor_batch=non_tensor_batch, meta_info=sub_meta_info) | |
| def index_select(self, index: Union[list[int], NDArray, torch.Tensor]) -> "DataProto": | |
| """Select a subset of the DataProto via index. | |
| Args: | |
| index (list, ndarray, torch.Tensor): a list of indices to select. | |
| Returns: | |
| DataProto: the DataProto containing the selected indices. | |
| """ | |
| if isinstance(index, list): | |
| index = np.array(index, dtype=bool if isinstance(index[0], bool) else np.int32) | |
| elif isinstance(index, torch.Tensor): | |
| index = index.detach().cpu().numpy() | |
| tensor_data = self.batch[index] if self.batch is not None else None | |
| non_tensor_data = {key: value[index] for key, value in self.non_tensor_batch.items()} | |
| return DataProto(batch=tensor_data, non_tensor_batch=non_tensor_data, meta_info=self.meta_info) | |
| def slice_select( | |
| self, start: Optional[int] = None, end: Optional[int] = None, step: Optional[int] = None | |
| ) -> "DataProto": | |
| """Select a subset of the DataProto via slice. | |
| Args: | |
| start (int, optional): the start index of the slice. | |
| end (int, optional): the end index of the slice. | |
| step (int, optional): the step of the slice. | |
| Returns: | |
| DataProto: the DataProto containing the selected slice. | |
| """ | |
| index = slice(start, end, step) | |
| tensor_data = self.batch[index] if self.batch is not None else None | |
| non_tensor_data = {key: value[index] for key, value in self.non_tensor_batch.items()} | |
| return DataProto(batch=tensor_data, non_tensor_batch=non_tensor_data, meta_info=self.meta_info) | |
| def pop( | |
| self, | |
| batch_keys: Optional[list[str]] = None, | |
| non_tensor_batch_keys: Optional[list[str]] = None, | |
| meta_info_keys: Optional[list[str]] = None, | |
| ) -> "DataProto": | |
| """Pop a subset of the DataProto via `batch_keys` and `meta_info_keys` | |
| Args: | |
| batch_keys (list, optional): a list of strings indicating the keys in batch to pop | |
| meta_info_keys (list, optional): a list of keys indicating the meta info to pop | |
| Returns: | |
| DataProto: the DataProto with the poped batch_keys and meta_info_keys | |
| """ | |
| assert batch_keys is not None | |
| non_tensor_batch_keys = non_tensor_batch_keys or [] | |
| meta_info_keys = meta_info_keys or [] | |
| tensors = {} | |
| for key in filter(lambda k: k in self.batch, batch_keys): | |
| tensors[key] = self.batch.pop(key) | |
| non_tensors = {} | |
| for key in filter(lambda k: k in self.non_tensor_batch, non_tensor_batch_keys): | |
| non_tensors[key] = self.non_tensor_batch.pop(key) | |
| meta_info = {} | |
| for key in filter(lambda k: k in self.meta_info, meta_info_keys): | |
| meta_info[key] = self.meta_info.pop(key) | |
| return DataProto.from_dict(tensors=tensors, non_tensors=non_tensors, meta_info=meta_info) | |
| def rename( | |
| self, old_keys: Optional[Union[str, list[str]]] = None, new_keys: Optional[Union[str, list[str]]] = None | |
| ) -> "DataProto": | |
| """ | |
| Note that this function only rename the key in the batch | |
| """ | |
| def validate_input(keys): | |
| if keys is not None: | |
| if isinstance(keys, str): | |
| keys = [keys] | |
| elif isinstance(keys, list): | |
| pass | |
| else: | |
| raise TypeError(f"keys must be a list or a string, but got {type(keys)}") | |
| return keys | |
| old_keys = validate_input(old_keys) | |
| new_keys = validate_input(new_keys) | |
| if len(new_keys) != len(old_keys): | |
| raise ValueError( | |
| f"new_keys and old_keys must have the same length, but got {len(new_keys)} and {len(old_keys)}" | |
| ) | |
| self.batch.rename_key_(tuple(old_keys), tuple(new_keys)) | |
| return self | |
| def union(self, other: "DataProto") -> "DataProto": | |
| """Union with another DataProto. Union batch and meta_info separately. | |
| Throw an error if | |
| - there are conflict keys in batch and they are not equal | |
| - the batch size of two data batch is not the same | |
| - there are conflict keys in meta_info and they are not the same. | |
| Args: | |
| other (DataProto): another DataProto to union | |
| Returns: | |
| DataProto: the DataProto after union | |
| """ | |
| self.batch = union_tensor_dict(self.batch, other.batch) | |
| self.non_tensor_batch = union_numpy_dict(self.non_tensor_batch, other.non_tensor_batch) | |
| self.meta_info = union_two_dict(self.meta_info, other.meta_info) | |
| return self | |
| def make_iterator( | |
| self, mini_batch_size: int, epochs: int, seed: int = None, dataloader_kwargs: dict[str, Any] = None | |
| ): | |
| """Make an iterator from the DataProto. This is built upon that TensorDict can be used as a normal Pytorch | |
| dataset. See https://pytorch.org/tensordict/tutorials/data_fashion for more details. | |
| Args: | |
| mini_batch_size (int): mini-batch size when iterating the dataset. We require that | |
| ``batch.batch_size[0] % mini_batch_size == 0`` | |
| epochs (int): number of epochs when iterating the dataset. | |
| dataloader_kwargs: internally, it returns a DataLoader over the batch. | |
| The dataloader_kwargs is the kwargs passed to the DataLoader | |
| Returns: | |
| Iterator: an iterator that yields a mini-batch data at a time. The total number of iteration steps is | |
| ``self.batch.batch_size * epochs // mini_batch_size`` | |
| """ | |
| assert self.batch.batch_size[0] % mini_batch_size == 0, f"{self.batch.batch_size[0]} % {mini_batch_size} != 0" | |
| if seed is not None: | |
| generator = torch.Generator() | |
| generator.manual_seed(seed) | |
| else: | |
| generator = None | |
| dataloader_kwargs = dataloader_kwargs or {} | |
| assert isinstance(dataloader_kwargs, dict) | |
| train_dataloader = DataLoader( | |
| dataset=self, | |
| batch_size=mini_batch_size, | |
| collate_fn=collate_fn, | |
| generator=generator, | |
| **dataloader_kwargs, | |
| ) | |
| def get_data(): | |
| for _ in range(epochs): | |
| for data in train_dataloader: | |
| setattr(data, "meta_info", self.meta_info) | |
| yield data | |
| return iter(get_data()) | |
| def chunk(self, chunks: int) -> list["DataProto"]: | |
| """Split the batch among dim=0 into chunks. The meta_info is passed to each DataProto after split. | |
| Args: | |
| chunks (int): the number of chunks to split on dim=0 | |
| Returns: | |
| List[DataProto]: a list of DataProto after splitting | |
| """ | |
| assert len(self) % chunks == 0, ( | |
| f"only support equal chunk. Got size of DataProto {len(self)} and chunk {chunks}." | |
| ) | |
| if self.batch is not None: | |
| batch_lst = self.batch.chunk(chunks=chunks, dim=0) | |
| else: | |
| batch_lst = [None for _ in range(chunks)] | |
| non_tensor_batch_lst = [{} for _ in range(chunks)] | |
| for key, value in self.non_tensor_batch.items(): | |
| non_tensor_lst = np.array_split(value, chunks) | |
| for i in range(chunks): | |
| non_tensor_batch_lst[i][key] = non_tensor_lst[i] | |
| return [ | |
| DataProto(batch=batch_lst[i], non_tensor_batch=non_tensor_batch_lst[i], meta_info=self.meta_info) | |
| for i in range(chunks) | |
| ] | |
| def split(self, split_size: int) -> list["DataProto"]: | |
| """Split the batch among dim=0 into chunks. The meta_info is passed to each DataProto after split. | |
| Args: | |
| split_size (int): the size of each split | |
| Returns: | |
| List[DataProto]: a list of DataProto after splitting | |
| """ | |
| assert len(self) % split_size == 0, ( | |
| f"only support equal split. Got size of DataProto {len(self)} and split {split_size}." | |
| ) | |
| chunks = len(self) // split_size | |
| return self.chunk(chunks) | |
| def concat(data: list["DataProto"]) -> "DataProto": | |
| """Concat a list of DataProto. The batch is concatenated among dim=0. | |
| The meta_info is assumed to be identical and will use the first one. | |
| Args: | |
| data (List[DataProto]): list of DataProto | |
| Returns: | |
| DataProto: concatenated DataProto | |
| """ | |
| batch_lst = [batch.batch for batch in data] | |
| new_batch = torch.cat(batch_lst, dim=0) if batch_lst[0] is not None else None | |
| non_tensor_batch = batch_collate([d.non_tensor_batch for d in data]) | |
| for key, value in non_tensor_batch.items(): | |
| non_tensor_batch[key] = np.concatenate(value, axis=0) | |
| return DataProto(batch=new_batch, non_tensor_batch=non_tensor_batch, meta_info=data[0].meta_info) | |
| def reorder(self, indices: torch.Tensor) -> None: | |
| """ | |
| Note that this operation is in-place | |
| """ | |
| indices_np = indices.detach().numpy() | |
| self.batch = self.batch[indices] | |
| self.non_tensor_batch = {key: value[indices_np] for key, value in self.non_tensor_batch.items()} | |
| def repeat(self, repeat_times: int, interleave: bool = True) -> "DataProto": | |
| """ | |
| Repeat the batch data a specified number of times. | |
| Args: | |
| repeat_times (int): Number of times to repeat the data. | |
| interleave (bool): Whether to interleave the repeated data. | |
| Returns: | |
| DataProto: A new DataProto with repeated data. | |
| """ | |
| if self.batch is not None: | |
| if interleave: # interleave the data | |
| repeated_tensors = { | |
| key: tensor.repeat_interleave(repeat_times, dim=0) for key, tensor in self.batch.items() | |
| } | |
| else: # stack the data | |
| repeated_tensors = { | |
| key: tensor.unsqueeze(0).expand(repeat_times, *tensor.shape).reshape(-1, *tensor.shape[1:]) | |
| for key, tensor in self.batch.items() | |
| } | |
| repeated_batch = TensorDict( | |
| source=repeated_tensors, | |
| batch_size=(self.batch.batch_size[0] * repeat_times,), | |
| ) | |
| else: | |
| repeated_batch = None | |
| repeated_non_tensor_batch = {} | |
| for key, value in self.non_tensor_batch.items(): | |
| if interleave: | |
| repeated_non_tensor_batch[key] = np.repeat(value, repeat_times, axis=0) | |
| else: | |
| repeated_non_tensor_batch[key] = np.tile(value, (repeat_times,) + (1,) * (value.ndim - 1)) | |
| return DataProto( | |
| batch=repeated_batch, | |
| non_tensor_batch=repeated_non_tensor_batch, | |
| meta_info=self.meta_info, | |
| ) | |
| class DataProtoFuture: | |
| """ | |
| DataProtoFuture aims to eliminate actual data fetching on driver. By doing so, the driver doesn't have to wait | |
| for data so that asynchronous execution becomes possible. | |
| DataProtoFuture contains a list of futures from another WorkerGroup of size world_size. | |
| - collect_fn is a Callable that reduces the list of futures to a DataProto | |
| - dispatch_fn is a Callable that partitions the DataProto into a list of DataProto of size world_size and then select | |
| Potential issue: we can optimize dispatch_fn(collect_fn) such that only needed data is fetched on destination | |
| - DataProtoFuture only supports directly passing from the output of a method to another input. You can't perform any | |
| operation on the DataProtoFuture in driver. | |
| """ | |
| collect_fn: Callable | |
| futures: list[ray.ObjectRef] | |
| dispatch_fn: Callable = None | |
| def concat(data: list[ray.ObjectRef]) -> "DataProtoFuture": | |
| output = DataProtoFuture(collect_fn=DataProto.concat, futures=data) | |
| return output | |
| def chunk(self, chunks: int) -> list["DataProtoFuture"]: | |
| from functools import partial | |
| arg_future_lst = [] | |
| for i in range(chunks): | |
| # note that we can't directly pass i and chunks | |
| def dispatch_fn(x, i, chunks): | |
| return x.chunk(chunks=chunks)[i] | |
| arg_future = DataProtoFuture( | |
| collect_fn=self.collect_fn, dispatch_fn=partial(dispatch_fn, i=i, chunks=chunks), futures=self.futures | |
| ) | |
| arg_future_lst.append(arg_future) | |
| return arg_future_lst | |
| def get(self): | |
| outputs = ray.get(self.futures) # dp_size | |
| for output in outputs: | |
| assert isinstance(output, DataProto) | |
| outputs = self.collect_fn(outputs) # select dp, concat | |
| if self.dispatch_fn is not None: | |
| outputs = self.dispatch_fn(outputs) # split in batch dim, select using dp | |
| return outputs | |
| def allgather_dict_tensors( | |
| tensors: Union[dict[str, torch.Tensor], TensorDict], size: int, group: ProcessGroup, dim: int = 0 | |
| ) -> Union[dict[str, torch.Tensor], TensorDict]: | |
| """ | |
| TODO: optimize this. | |
| - We can use async ops | |
| - We can use only one allgather | |
| """ | |
| if isinstance(tensors, TensorDict): | |
| is_tensor_dict = True | |
| tensors_as_dict = tensors.to_dict() | |
| else: | |
| tensors_as_dict = tensors | |
| is_tensor_dict = False | |
| output = {} | |
| sorted_keys = sorted(tensors_as_dict.keys()) | |
| for key in sorted_keys: | |
| value = tensors_as_dict[key] | |
| output[key] = [torch.empty_like(value) for _ in range(size)] | |
| torch.distributed.all_gather(output[key], value, group=group, async_op=False) | |
| output[key] = torch.cat(output[key], dim=dim) | |
| if is_tensor_dict: | |
| output = TensorDict(source=output, batch_size=tensors.batch_size[0] * size) | |
| return output | |
| def all_gather_data_proto(data: DataProto, size: int, group: ProcessGroup) -> None: | |
| # Note that this is an inplace operator just like torch.distributed.all_gather | |
| prev_device = data.batch.device | |
| data.batch = data.batch.cuda(device=torch.cuda.current_device()) | |
| data.batch = allgather_dict_tensors(data.batch.contiguous(), size=size, group=group, dim=0) | |
| data.batch = data.batch.to(prev_device) | |
| # all gather non_tensor_batch | |
| all_non_tensor_batch = [None for _ in range(size)] | |
| torch.distributed.all_gather_object(all_non_tensor_batch, data.non_tensor_batch, group=group) | |
| data.non_tensor_batch = {k: np.concatenate([d[k] for d in all_non_tensor_batch]) for k in data.non_tensor_batch} | |