Ng ddlmZddlZddlmZddlmZddlZddlZddl m Z ddlm Z m Z ddl mZejeZGdd e jZdS) ) annotationsN)Iterable)Callable)version)Tensornn)SentenceTransformercdeZdZdddejfdfd ZddZeddZxZ S) SoftmaxLossTFmodelr sentence_embedding_dimensionint num_labelsconcatenation_sent_repboolconcatenation_sent_difference!concatenation_sent_multiplicationloss_fctrreturnNonect||_||_||_||_||_d}|r|dz }|r|dz }|r|dz }td|tj ||z||j |_ ||_ tjt jtjdkrtddSdS) a\ This loss was used in our SBERT publication (https://arxiv.org/abs/1908.10084) to train the SentenceTransformer model on NLI data. It adds a softmax classifier on top of the output of two transformer networks. :class:`MultipleNegativesRankingLoss` is an alternative loss function that often yields better results, as per https://arxiv.org/abs/2004.09813. Args: model (SentenceTransformer): The SentenceTransformer model. sentence_embedding_dimension (int): The dimension of the sentence embeddings. num_labels (int): The number of different labels. concatenation_sent_rep (bool): Whether to concatenate vectors u,v for the softmax classifier. Defaults to True. concatenation_sent_difference (bool): Whether to add abs(u-v) for the softmax classifier. Defaults to True. concatenation_sent_multiplication (bool): Whether to add u*v for the softmax classifier. Defaults to False. loss_fct (Callable): Custom pytorch loss function. If not set, uses nn.CrossEntropyLoss(). Defaults to nn.CrossEntropyLoss(). References: - Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks: https://arxiv.org/abs/1908.10084 - `Training Examples > Natural Language Inference <../../examples/training/nli/README.html>`_ Requirements: 1. sentence pairs with a class label Inputs: +---------------------------------------+--------+ | Texts | Labels | +=======================================+========+ | (sentence_A, sentence_B) pairs | class | +---------------------------------------+--------+ Example: :: from sentence_transformers import SentenceTransformer, SentenceTransformerTrainer, losses from datasets import Dataset model = SentenceTransformer("microsoft/mpnet-base") train_dataset = Dataset.from_dict({ "sentence1": [ "A person on a horse jumps over a broken down airplane.", "A person on a horse jumps over a broken down airplane.", "A person on a horse jumps over a broken down airplane.", "Children smiling and waving at camera", ], "sentence2": [ "A person is training his horse for a competition.", "A person is at a diner, ordering an omelette.", "A person is outdoors, on a horse.", "There are children present.", ], "label": [1, 2, 0, 0], }) loss = losses.SoftmaxLoss(model, model.get_sentence_embedding_dimension(), num_labels=3) trainer = SentenceTransformerTrainer( model=model, train_dataset=train_dataset, loss=loss, ) trainer.train() rz%Softmax loss: #Vectors concatenated: )devicez4.43.0zSoftmaxLoss requires transformers >= 4.43.0 to work correctly. Otherwise, the classifier layer that maps embeddings to the labels cannot be updated. Consider updating transformers with `pip install transformers>=4.43.0`.N)super__init__r rrrrloggerinforLinearr classifierrrparse transformers __version__warning) selfr r rrrrrnum_vectors_concatenated __class__s d/var/www/html/ai-engine/env/lib/python3.11/site-packages/sentence_transformers/losses/SoftmaxLoss.pyrzSoftmaxLoss.__init__s N  $&<#-J*1R.#$ ! * $ ) $ ( * $ ) $ , * $ ) $ Vz'SoftmaxLoss.forward..ws-mmmGW +,,-ABmmmr)r) rappendrtorchabsrcatr rview) r%r*r,repsrep_arep_bvectors_concatfeaturesoutputlosss ` r(forwardzSoftmaxLoss.forwardtsnmmm[lmmm u  & )  ! !% ( ( (  ! !% ( ( (  - <  ! !%)EEM":": ; ; ;  1 1  ! !%%- 0 0 09^Q//**  ==R99DK< r)strcdS)Na @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "https://arxiv.org/abs/1908.10084", } )r%s r(citationzSoftmaxLoss.citations   r))r r r rrrrrrrrrrrrr)r*r+r,rrr-)rrB) __name__ __module__ __qualname__rCrossEntropyLossrrApropertyrE __classcell__)r's@r(r r s (,.2270R022```````D    6   X     r)r ) __future__rloggingcollections.abcrtypingrr6r" packagingrrr)sentence_transformers.SentenceTransformerr getLoggerrFrModuler rDr)r(rTs""""""$$$$$$ IIIIII  8 $ $JJJJJ")JJJJJr)