o
    shS                     @   st   d dl mZ d dlmZmZmZ d dlm  mZ	 d dl
mZmZ d dlmZ G dd deZG dd	 d	ejZdS )
    )Enum)AnyDictIterableN)Tensornn)SentenceTransformerc                   @   s(   e Zd ZdZdd Zdd Zdd ZdS )SiameseDistanceMetricz#The metric for the contrastive lossc                 C      t j| |ddS )N   pFpairwise_distancexy r   j/var/www/html/alpaca_bot/venv/lib/python3.10/site-packages/sentence_transformers/losses/ContrastiveLoss.py<lambda>       zSiameseDistanceMetric.<lambda>c                 C   r
   )N   r   r   r   r   r   r   r      r   c                 C   s   dt | | S )Nr   )r   cosine_similarityr   r   r   r   r      r   N)__name__
__module____qualname____doc__	EUCLIDEAN	MANHATTANCOSINE_DISTANCEr   r   r   r   r	   
   s
    r	   c                	       s   e Zd Zejddfdedededdf fdd	Zde	e
ef fd
dZdee	e
ef  dedefddZede
fddZ  ZS )ContrastiveLoss      ?Tmodelmarginsize_averagereturnNc                    s*   t t|   || _|| _|| _|| _dS )a	  
        Contrastive loss. Expects as input two texts and a label of either 0 or 1. If the label == 1, then the distance between the
        two embeddings is reduced. If the label == 0, then the distance between the embeddings is increased.

        Args:
            model: SentenceTransformer model
            distance_metric: Function that returns a distance between
                two embeddings. The class SiameseDistanceMetric contains
                pre-defined metrices that can be used
            margin: Negative samples (label == 0) should have a distance
                of at least the margin value.
            size_average: Average by the size of the mini-batch.

        References:
            * Further information: http://yann.lecun.com/exdb/publis/pdf/hadsell-chopra-lecun-06.pdf
            * `Training Examples > Quora Duplicate Questions <../../examples/training/quora_duplicate_questions/README.html>`_

        Requirements:
            1. (anchor, positive/negative) pairs

        Relations:
            - :class:`OnlineContrastiveLoss` is similar, but uses hard positive and hard negative pairs.
            It often yields better results.

        Inputs:
            +-----------------------------------------------+------------------------------+
            | Texts                                         | Labels                       |
            +===============================================+==============================+
            | (anchor, positive/negative) pairs             | 1 if positive, 0 if negative |
            +-----------------------------------------------+------------------------------+

        Example:
            ::

                from sentence_transformers import SentenceTransformer, SentenceTransformerTrainer, losses
                from datasets import Dataset

                model = SentenceTransformer("microsoft/mpnet-base")
                train_dataset = Dataset.from_dict({
                    "sentence1": ["It's nice weather outside today.", "He drove to work."],
                    "sentence2": ["It's so sunny.", "She walked to the store."],
                    "label": [1, 0],
                })
                loss = losses.ContrastiveLoss(model)

                trainer = SentenceTransformerTrainer(
                    model=model,
                    train_dataset=train_dataset,
                    loss=loss,
                )
                trainer.train()
        N)superr!   __init__distance_metricr$   r#   r%   )selfr#   r)   r$   r%   	__class__r   r   r(      s
   ;
zContrastiveLoss.__init__c                 C   sF   | j j}tt D ]\}}|| j krd|} nq
|| j| jdS )NzSiameseDistanceMetric.{})r)   r$   r%   )r)   r   varsr	   itemsformatr$   r%   )r*   distance_metric_namenamevaluer   r   r   get_config_dictT   s   

zContrastiveLoss.get_config_dictsentence_featureslabelsc                    s    fdd|D }t |dksJ |\}} ||}d| |d d|  t j| d   } jr>| S |	 S )Nc                    s   g | ]	}  |d  qS )sentence_embedding)r#   ).0sentence_featurer*   r   r   
<listcomp>^   s    z+ContrastiveLoss.forward.<locals>.<listcomp>r   r"   r   )
lenr)   floatpowr   relur$   r%   meansum)r*   r4   r5   reps
rep_anchor	rep_other	distanceslossesr   r9   r   forward]   s   2zContrastiveLoss.forwardc                 C   s   dS )Na  
@inproceedings{hadsell2006dimensionality,
    author={Hadsell, R. and Chopra, S. and LeCun, Y.},
    booktitle={2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06)}, 
    title={Dimensionality Reduction by Learning an Invariant Mapping}, 
    year={2006},
    volume={2},
    number={},
    pages={1735-1742},
    doi={10.1109/CVPR.2006.100}
}
r   r9   r   r   r   citationg   s   zContrastiveLoss.citation)r   r   r   r	   r    r   r<   boolr(   r   strr   r3   r   r   rF   propertyrG   __classcell__r   r   r+   r   r!      s"    A"	
r!   )enumr   typingr   r   r   torch.nn.functionalr   
functionalr   torchr   )sentence_transformers.SentenceTransformerr   r	   Moduler!   r   r   r   r   <module>   s    