PiyddlmZmZmZmZmZddlmZddlm Z ddl m Z ddl m Z ddddddd d d e d ed eeeefdedeededeededeeefdee e ffdZdS))AnyCallableDictOptionalUnion)InputOutputToMessages) PackedDataset) SFTDataset)ModelTokenizerzSamsung/samsumNFtrain)source column_maptrain_on_inputnew_system_promptpacked filter_fnsplit tokenizerr rrrrrrload_dataset_kwargsreturnc |pddd}t|||} td || |||d|} |r,|jtdt | |jS| S) a+ Support for summarization datasets and their variants from Hugging Face Datasets. An example is the `SAMsum dataset `_. It is recommended to configure the tokenizer with the :class:`~torchtune.data.SummarizeTemplate` in conjunction with this dataset. Masking of the prompt during training is controlled by the ``train_on_input`` flag, which is set to ``False`` by default - If ``train_on_input`` is True, the prompt is used during training and contributes to the loss. - If ``train_on_input`` is False, the prompt is masked out (tokens replaced with -100) Args: tokenizer (ModelTokenizer): Tokenizer used by the model that implements the ``tokenize_messages`` method. source (str): path to dataset repository on Hugging Face. For local datasets, define source as the data file type (e.g. "json", "csv", "text"), pass in the filepath in ``data_files``, and set ``split="train"``. See `Hugging Face's `_ ``load_dataset`` for more details. Default is ``Samsung/samsum``. column_map (Optional[Dict[str, str]]): a mapping from the expected columns in the message transform :class:`~torchtune.data.InputOutputToMessages` to the new column names in the dataset. Keys should be "input" and "output" and values should be the actual column names. If None, use the default column names ``{"input": "dialogue", "output": "summary"}`` in ``Samsung/samsum``. train_on_input (bool): Whether the model is trained on the prompt or not. Default is False. new_system_prompt (Optional[str]): if specified, prepend a system message to every sample. This can serve as instructions to guide the model response. Setting this will OVERRIDE any system messages already present in the dataset. Default is None. packed (bool): Whether or not to pack the dataset to tokenizer's ``max_seq_len`` prior to training. Default is False. filter_fn (Optional[Callable]): callable used to filter the dataset prior to any pre-processing. See the Hugging Face `docs `_ for more details. split (str): ``split`` argument for ``datasets.load_dataset``. You can use this argument to load a subset of a given split, e.g. ``split="train[:10%]"``. Default is "train". **load_dataset_kwargs (Dict[str, Any]): additional keyword arguments to pass to ``load_dataset``. Returns: Union[SFTDataset, PackedDataset]: dataset configured with source data and template Raises: ValueError: If ``packed=True`` and ``tokenizer.max_seq_len`` is not set. Example: >>> samsum_ds = samsum_dataset(model_transform=tokenizer) >>> for batch in Dataloader(samsum_ds, batch_size=8): >>> print(f"Batch size: {len(batch)}") >>> Batch size: 8 dialoguesummary)inputoutput)rrr)r message_transformmodel_transformrrNz@PackedDataset requires a max_seq_len to be set on the tokenizer.) max_seq_len)rr r ValueErrorr ) rr rrrrrrrrdss n/home/jaya/work/projects/VOICE-AGENT/VIET/agent-env/lib/python3.11/site-packages/torchtune/datasets/_samsum.pysamsum_datasetr#sxIyIIJ-%+  +!       BD  (R RY-BCCCC I)typingrrrrrtorchtune.datartorchtune.datasets._packedr torchtune.datasets._sftr 'torchtune.modules.transforms.tokenizersr strboolr#rr$r"r,sT87777777777777000000444444......BBBBBB #+/ '+$(QQQQ Qc3h( Q  Q  } Q Q!Q Q S>Q :} $%QQQQQQr$