o sh@sddlZddlmZmZddlmZddlmZddl m Z ddl m Z e eZGddde ZGd d d e ZeGd d d eZdS) N) dataclassfield)Union)TrainingArguments) ParallelMode) ExplicitEnumc@seZdZdZdZdZdZdS) BatchSamplersa Stores the acceptable string identifiers for batch samplers. The batch sampler is responsible for determining how samples are grouped into batches during training. Valid options are: - ``BatchSamplers.BATCH_SAMPLER``: The default PyTorch batch sampler. - ``BatchSamplers.NO_DUPLICATES``: Ensures no duplicate samples in a batch. - ``BatchSamplers.GROUP_BY_LABEL``: Ensures each batch has 2+ samples from the same label. batch_sampler no_duplicatesgroup_by_labelN)__name__ __module__ __qualname____doc__ BATCH_SAMPLER NO_DUPLICATESGROUP_BY_LABELrra/var/www/html/alpaca_bot/venv/lib/python3.10/site-packages/sentence_transformers/training_args.pyr s  rc@seZdZdZdZdZdS)MultiDatasetBatchSamplersa Stores the acceptable string identifiers for multi-dataset batch samplers. The multi-dataset batch sampler is responsible for determining in what order batches are sampled from multiple datasets during training. Valid options are: - ``MultiDatasetBatchSamplers.ROUND_ROBIN``: Round-robin sampling from each dataset until one is exhausted. With this strategy, it's likely that not all samples from each dataset are used, but each dataset is sampled from equally. - ``MultiDatasetBatchSamplers.PROPORTIONAL``: Sample from each dataset in proportion to its size [default]. With this strategy, all samples from each dataset are used and larger datasets are sampled from more frequently. round_robin proportionalN)r r rr ROUND_ROBIN PROPORTIONALrrrrrs rcsfeZdZUdZeejddidZeee fe d<ee j ddidZ ee e fe d<fdd ZZS) $SentenceTransformerTrainingArgumentsa SentenceTransformerTrainingArguments extends :class:`~transformers.TrainingArguments` with additional arguments specific to Sentence Transformers. See :class:`~transformers.TrainingArguments` for the complete list of available arguments. Args: output_dir (`str`): The output directory where the model checkpoints will be written. batch_sampler (Union[:class:`~sentence_transformers.training_args.BatchSamplers`, `str`], *optional*): The batch sampler to use. See :class:`~sentence_transformers.training_args.BatchSamplers` for valid options. Defaults to ``BatchSamplers.BATCH_SAMPLER``. multi_dataset_batch_sampler (Union[:class:`~sentence_transformers.training_args.MultiDatasetBatchSamplers`, `str`], *optional*): The multi-dataset batch sampler to use. See :class:`~sentence_transformers.training_args.MultiDatasetBatchSamplers` for valid options. Defaults to ``MultiDatasetBatchSamplers.PROPORTIONAL``. helpzThe batch sampler to use.)defaultmetadatar z'The multi-dataset batch sampler to use.multi_dataset_batch_samplercstt|j|_t|j|_d|_d|_|jt j kr+|j dkr)t ddSdS|jt jkrC|jsE|j dkr>t dd|_dSdSdS)NTFunusedzCurrently using DataParallel (DP) for multi-gpu training, while DistributedDataParallel (DDP) is recommended for faster training. See https://sbert.net/docs/sentence_transformer/training/distributed.html for more information.zWhen using DistributedDataParallel (DDP), it is recommended to set `dataloader_drop_last=True` to avoid hanging issues with an uneven last batch. Setting `dataloader_drop_last=True`.)super __post_init__rr rrprediction_loss_onlyddp_broadcast_buffers parallel_moderNOT_DISTRIBUTED output_dirloggerwarning DISTRIBUTEDdataloader_drop_last)self __class__rrr!Hs$       z2SentenceTransformerTrainingArguments.__post_init__)r r rrrrrr rstr__annotations__rrrr! __classcell__rrr,rr/s   r)logging dataclassesrrtypingr transformersrTransformersTrainingArgumentstransformers.training_argsrtransformers.utilsr getLoggerr r'rrrrrrrs