§ ÇPƒišãóÜ—ddlmZmZmZmZmZddlmZddlm Z ddl m Z ddl m Z Gd„de ¦«Zdd d d d d œd e dedeeeefdedeededeeefde fd„Zd S)é)ÚAnyÚCallableÚDictÚMappingÚOptional)ÚMessage)ÚPreferenceDataset)Ú Transform)ÚModelTokenizercór—eZdZdZ d dedeeeeffd„Zde ee fde ee ffd „Z dS) ÚStackExchangePairedToMessagesaj Transform for converting datasets similar to the format in `Stack Exchange Paired dataset `_:: | prompt | chosen | rejected | |----------|----------|------------| | Q1 | A1 | A2 | into a list of chosen and rejected messages: .. code-block:: python chosen = [ Message(role="user", content="Q1"), Message(role="assistant", content="A1"), ] rejected = [ Message(role="user", content="Q1"), Message(role="assistant", content="A2"), ] Args: train_on_input (bool): Whether the model is trained on the user prompt or not. Default is False. column_map (Optional[Dict[str, str]]): a mapping to change the expected "prompt", "chosen", and "rejected" column names to the actual column names in the dataset. Keys should be "prompt", "chosen", and "rejected" and values should be the actual column names. Default is None, keeping the default column names. FNÚtrain_on_inputÚ column_mapcó"—||_||_dS)N)rÚ _column_map)Úselfrrs ú}/home/jaya/work/projects/VOICE-AGENT/VIET/agent-env/lib/python3.11/site-packages/torchtune/datasets/_stack_exchange_paired.pyÚ__init__z&StackExchangePairedToMessages.__init__.s€ð-ˆÔØ%ˆÔÐÐóÚsampleÚreturncóv—|jpi}| dd¦«}| dd¦«}| dd¦«}td|||j ¬¦«td||¬¦«g}td|||j ¬¦«td||¬¦«g}||dœS) NÚpromptÚchosenÚrejectedÚuser)ÚroleÚcontentÚmaskedÚ assistant)rr)rr)rÚgetrr)rrrÚ key_promptÚ key_chosenÚ key_rejectedÚchosen_messagesÚrejected_messagess rÚ__call__z&StackExchangePairedToMessages.__call__4sâ€ØÔ%Ð+¨ˆ Ø—^’^ H¨hÑ7Ô7ˆ Ø—^’^ H¨hÑ7Ô7ˆ Ø!—~’~ j°*Ñ=Ô=ˆ õ Ø V¨JÔ%7ÀDÔDWÐ@Wð ñ ô õ ˜¨f°ZÔ.@Ð AÑ AÔ Að  ˆõ Ø V¨JÔ%7ÀDÔDWÐ@Wð ñ ô õ ˜¨f°\Ô.BÐ CÑ CÔ Cð  Ðð*Ð7HÐIÐIÐIr)FN) Ú__name__Ú __module__Ú __qualname__Ú__doc__ÚboolrrÚstrrrrr'©rrr r s—€€€€€ððð>TXð&ð&Ø"ð&Ø8@ÀÀcÈ3ÀhÄÔ8Pð&ð&ð&ð&ð J˜w s¨C xÔ0ðJ°W¸SÀ#¸XÔ5FðJðJðJðJðJðJrr zlvwerra/stack-exchange-pairedNFÚtrain)ÚsourcerrÚ filter_fnÚsplitÚ tokenizerr0rrr1r2Úload_dataset_kwargsrc óZ—|pddddœ}t||¬¦«}td|||||ddœ|¤ŽS) aµ Family of preference datasets similar to the `Stack Exchange Paired dataset `_. It is recommended to configure the tokenizer with the :class:`~torchtune.data.QuestionAnswerTemplate` in conjunction with this dataset. 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") and pass in the filepath in ``data_files``. See `Hugging Face's `_ ``load_dataset`` for more details. Default is ``lvwerra/stack-exchange-paired``. column_map (Optional[Dict[str, str]]): a mapping to change the expected "prompt", "chosen", and "rejected" column names to the actual column names in the dataset. Keys should be "prompt", "chosen", and "rejected" and values should be the actual column names. Default is None, keeping the default column names. train_on_input (bool): Whether the model is trained on the prompt or not. 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: PreferenceDataset: The preference dataset built from source paired data. ÚquestionÚ response_jÚ response_k)rrr)rrzdata/rl)r0Úmessage_transformr3r1r2Údata_dirr.)r r )r3r0rrr1r2r4r9s rÚstack_exchange_paired_datasetr;Ksx€ðPðØØØ ð ð €Jõ 6Ø%°*ðñôÐõ ð ØØ+ØØØØð  ð ð ð ð ðr)ÚtypingrrrrrÚtorchtune.datarÚtorchtune.datasets._preferencer Útorchtune.modules.transformsr Ú'torchtune.modules.transforms.tokenizersr r r-r,r;r.rrúrAsMðð:Ð9Ð9Ð9Ð9Ð9Ð9Ð9Ð9Ð9Ð9Ð9Ð9Ð9à"Ð"Ð"Ð"Ð"Ð"Ø<Ð<Ð<Ð<Ð<Ð<Ø2Ð2Ð2Ð2Ð2Ð2ØBÐBÐBÐBÐBÐBð9Jð9Jð9Jð9Jð9J Iñ9Jô9Jð9Jð~2Ø+/Ø Ø$(Øð:ð:ð:Øð:ð ð:ð˜˜c 3˜hœÔ(ð :ð ð :ð ˜Ô!ð :ð ð:ð   S œ>ð:ðð:ð:ð:ð:ð:ð:r