a hR @sVddlmZddlZddlZddlmZmZddlmZee Z GdddeZ dS)) annotationsN)Tensornn)ModulecsdeZdZUdZgdZded<dddddfdd Zd d d d ZddddddddZZ S) WordWeightszDThis model can weight word embeddings, for example, with idf-values.)vocab word_weightsunknown_word_weightz list[str] config_keyszdict[str, float]floatcst||_||_||_g}d}|D]F}|}||vrB||}n"||vr\||}n|d7}||q(t|dt |d|t t |d|_ |j dt|didS)aZ Initializes the WordWeights class. Args: vocab (List[str]): Vocabulary of the tokenizer. word_weights (Dict[str, float]): Mapping of tokens to a float weight value. Word embeddings are multiplied by this float value. Tokens in word_weights must not be equal to the vocab (can contain more or less values). unknown_word_weight (float, optional): Weight for words in vocab that do not appear in the word_weights lookup. These can be, for example, rare words in the vocab where no weight exists. Defaults to 1. rr z of z0 words without a weighting value. Set weight to weightN)super__init__rrr lowerappendloggerinfolenrZ Embedding emb_layerZload_state_dicttorchZ FloatTensor unsqueeze)selfrrr weightsZnum_unknown_wordswordr  __class__d/var/www/html/assistant/venv/lib/python3.9/site-packages/sentence_transformers/models/WordWeights.pyrs&    zWordWeights.__init__zdict[str, Tensor])featurescCsl|d}|d}||dd}||}t|d}|d|}||}|||d|S)Nattention_masktoken_embeddingsZ input_idsr )r!token_weights_sum) rZsqueezer rsumrexpandsizeupdate)rrr r!Ztoken_weights_rawZ token_weightsr#Ztoken_weights_expandedrrrforward5s  zWordWeights.forwardT)safe_serializationstrboolNone) output_pathr)returncOs||dS)N)Z save_config)rr-r)argskwargsrrrsaveEszWordWeights.save)r ) __name__ __module__ __qualname____doc__r __annotations__rr(r1 __classcell__rrrrr s #r) __future__rloggingrrrZ#sentence_transformers.models.Moduler getLoggerr2rrrrrrs