rhddlmZddlZddlZddlZddlZddlmZddlm Z m Z m Z ddl Z ddlZddlmZddlmZddlmZddlmZdd lmZmZmZmZmZmZdd lmZdd lm Z dd l!m"Z"dd l#m$Z$m%Z%ddl&m'Z'ddl(m)Z)m*Z*ddl+m,Z,m-Z-ddl.m/Z/m0Z0m1Z1m2Z2ejfe4Z5ddZ6Gddejnee'Z8y)) annotationsN)Path)CallableLiteraloverload)HfApi)version)nn)trange) AutoConfig"AutoModelForSequenceClassification AutoTokenizerPretrainedConfigPreTrainedModelPreTrainedTokenizer)PushToHubMixin) deprecated) __version__)load_onnx_modelload_openvino_model)FitMixin)CrossEncoderModelCardDatagenerate_model_card)!cross_encoder_init_args_decorator)cross_encoder_predict_rank_args_decorator)fullnameget_device_nameimport_from_stringload_file_pathcdfd }|S)Ncttjt|z dt|z fi|S)NT)exist_ok)osmakedirsr)save_directorykwargs_save_pretrained_fn subfolders /var/www/html/ai-insurance-compliance-backend/venv/lib/python3.12/site-packages/sentence_transformers/cross_encoder/CrossEncoder.pywrapperz)_save_pretrained_wrapper..wrapper)s5 D(94tD"4#7)#CNvNN)r%z str | PathreturnNone)r'r(r*s`` r)_save_pretrained_wrapperr/(sO Nr+c eZdZdZe d# d$fd Z d%dZd&dZd'd(dZd)d Z d*d Z e d+d Z e d,d Z e d,d Zejd-dZe edd)dZdZe d. d/dZe d0 d1dZe d. d2dZe d. d3dZej,e d4 d5dZe d6 d7dZddd8dZddd8dZd9dZddddddddd d:dZe d;dZe dd Ze d 1, it output several scores that can be soft-maxed to get probability scores for the different classes. Defaults to None. max_length (int, optional): Max length for input sequences. Longer sequences will be truncated. If None, max length of the model will be used. Defaults to None. activation_fn (Callable, optional): Callable (like nn.Sigmoid) about the default activation function that should be used on-top of model.predict(). If None. nn.Sigmoid() will be used if num_labels=1, else nn.Identity(). Defaults to None. device (str, optional): Device (like "cuda", "cpu", "mps", "npu") that should be used for computation. If None, checks if a GPU can be used. cache_folder (`str`, `Path`, optional): Path to the folder where cached files are stored. trust_remote_code (bool, optional): Whether or not to allow for custom models defined on the Hub in their own modeling files. This option should only be set to True for repositories you trust and in which you have read the code, as it will execute code present on the Hub on your local machine. Defaults to False. revision (str, optional): The specific model version to use. It can be a branch name, a tag name, or a commit id, for a stored model on Hugging Face. Defaults to None. local_files_only (bool, optional): Whether or not to only look at local files (i.e., do not try to download the model). token (bool or str, optional): Hugging Face authentication token to download private models. model_kwargs (Dict[str, Any], optional): Additional model configuration parameters to be passed to the Hugging Face Transformers model. Particularly useful options are: - ``torch_dtype``: Override the default `torch.dtype` and load the model under a specific `dtype`. The different options are: 1. ``torch.float16``, ``torch.bfloat16`` or ``torch.float``: load in a specified ``dtype``, ignoring the model's ``config.torch_dtype`` if one exists. If not specified - the model will get loaded in ``torch.float`` (fp32). 2. ``"auto"`` - A ``torch_dtype`` entry in the ``config.json`` file of the model will be attempted to be used. If this entry isn't found then next check the ``dtype`` of the first weight in the checkpoint that's of a floating point type and use that as ``dtype``. This will load the model using the ``dtype`` it was saved in at the end of the training. It can't be used as an indicator of how the model was trained. Since it could be trained in one of half precision dtypes, but saved in fp32. - ``attn_implementation``: The attention implementation to use in the model (if relevant). Can be any of `"eager"` (manual implementation of the attention), `"sdpa"` (using `F.scaled_dot_product_attention `_), or `"flash_attention_2"` (using `Dao-AILab/flash-attention `_). By default, if available, SDPA will be used for torch>=2.1.1. The default is otherwise the manual `"eager"` implementation. See the `AutoModelForSequenceClassification.from_pretrained `_ documentation for more details. tokenizer_kwargs (Dict[str, Any], optional): Additional tokenizer configuration parameters to be passed to the Hugging Face Transformers tokenizer. See the `AutoTokenizer.from_pretrained `_ documentation for more details. config_kwargs (Dict[str, Any], optional): Additional model configuration parameters to be passed to the Hugging Face Transformers config. See the `AutoConfig.from_pretrained `_ documentation for more details. For example, you can set ``classifier_dropout`` via this parameter. model_card_data (:class:`~sentence_transformers.model_card.SentenceTransformerModelCardData`, optional): A model card data object that contains information about the model. This is used to generate a model card when saving the model. If not set, a default model card data object is created. backend (str): The backend to use for inference. Can be one of "torch" (default), "onnx", or "openvino". See https://sbert.net/docs/cross_encoder/usage/efficiency.html for benchmarking information on the different backends. NFc vt|| i} | i} | i} |xs t| |_||_d|_||_tj|f|||| | d| }t|drld|jvr^|jd}tj|tjtkDr tjd|dtdd}|j 2t#|j Dcgc]}|j%d c}}||sd }|||_|j(|f|||||| | d | d | vr||| d <t+j|f|||| | d| |_d | vrYt|j.d rCt1|j,j2|j.j4|j,_t7|d| j9dd||| }|, t;|d5}|j=|_ddd|jA||"tC}tjEd||jG||jjI||jjK||ycc}w#1swYxYw#t>$rYwxYw)N)local_files_only) cache_dirtrust_remote_coderevisionr3tokensentence_transformersr zRYou are trying to use a model that was created with Sentence Transformers version z%, but you're currently using version zd. This might cause unexpected behavior or errors. In that case, try to update to the latest version.FForSequenceClassification)configbackendr4r5r6r3r7model_max_lengthmax_position_embeddings README.mdr7)r7 cache_folderr6r3utf8encodingzUse pytorch device: )r6)&super__init__rmodel_card_datar5_model_card_textr<r from_pretrainedhasattrr8r parserloggerwarning architecturesanyendswith num_labels _load_modelr tokenizerr;minr=r>rgetopenread Exceptionset_activation_fnrinfotoregister_modelset_base_model)selfmodel_name_or_pathrP max_length activation_fndevicer@r5r6r3r7 model_kwargstokenizer_kwargs config_kwargsrFr<r; model_versionclassifier_trainedarchmodel_card_pathfIn __class__s r)rEzCrossEncoder.__init__vs &   #!   L  M.n2K]m2n!2 $ #-#=#= $ "/- $ $  62 3 VEaEa8a"88CM}}]+gmmK.HHhivhwx::EGII #    +!$]c]q]q%rUYdmm4O&P%r!s   &8J  ! *F   "/-   %5 5*:P3= / 0.;.K.K / "/- / /  %5 5'$++Oh:i.1$..2Q2QSWS^S^SvSv.wDNN +)  ""7D1%-    & /F;7s,/HHJD)7 }- >$&F KK.vh7 8  ++D1 ++,>+Ru&sX77  s0;J1 J,>J J, J)%J,, J87J8c |dk(rtj|fd|i||_y|dk(rtd ||dd||_y|dk(rt d ||dd||_yt d|d) Ntorchr;onnxzsequence-classification)r^r; task_nameopenvinozUnsupported backend 'z6'. `backend` should be `torch`, `onnx`, or `openvino`.r.)r rHmodelrr ValueError)r]r^r;r<rbs r)rQzCrossEncoder._load_models g *L*\*\"+++DJ  (#53 DJ  ",#53 DJ4WI=stu ur+c|jS)zReturn the backend used for inference, which can be one of "torch", "onnx", or "openvino". Returns: str: The backend used for inference. )r<r]s r) get_backendzCrossEncoder.get_backends ||r+Tc|||_n|j|_|r&|jdt|jyy)Nr`)r`get_default_activation_fnset_config_valuer)r]r` set_defaults r)rXzCrossEncoder.set_activation_fn sD  $!.D !%!?!?!AD    ! !/8Dc}} |r| d} | Scc}wcc}w)a Performs predictions with the CrossEncoder on the given sentence pairs. Args: sentences (Union[List[Tuple[str, str]], Tuple[str, str]]): A list of sentence pairs [(Sent1, Sent2), (Sent3, Sent4)] or one sentence pair (Sent1, Sent2). batch_size (int, optional): Batch size for encoding. Defaults to 32. show_progress_bar (bool, optional): Output progress bar. Defaults to None. activation_fn (callable, optional): Activation function applied on the logits output of the CrossEncoder. If None, the ``model.activation_fn`` will be used, which defaults to :class:`torch.nn.Sigmoid` if num_labels=1, else :class:`torch.nn.Identity`. Defaults to None. convert_to_numpy (bool, optional): Convert the output to a numpy matrix. Defaults to True. apply_softmax (bool, optional): If set to True and `model.num_labels > 1`, applies softmax on the logits output such that for each sample, the scores of each class sum to 1. Defaults to False. convert_to_numpy (bool, optional): Whether the output should be a list of numpy vectors. If False, output a list of PyTorch tensors. Defaults to True. convert_to_tensor (bool, optional): Whether the output should be one large tensor. Overwrites `convert_to_numpy`. Defaults to False. Returns: Union[List[torch.Tensor], np.ndarray, torch.Tensor]: Predictions for the passed sentence pairs. The return type depends on the ``convert_to_numpy`` and ``convert_to_tensor`` parameters. If ``convert_to_tensor`` is True, the output will be a :class:`torch.Tensor`. If ``convert_to_numpy`` is True, the output will be a :class:`numpy.ndarray`. Otherwise, the output will be a list of :class:`torch.Tensor` values. Examples: :: from sentence_transformers import CrossEncoder model = CrossEncoder("cross-encoder/stsb-roberta-base") sentences = [["I love cats", "Cats are amazing"], ["I prefer dogs", "Dogs are loyal"]] model.predict(sentences) # => array([0.6912767, 0.4303499], dtype=float32) FrT)rxBatches)descdisablept)padding truncationreturn_tensors return_dictr:)dim)rar.)# isinstancelisttuplerrKgetEffectiveLevelloggingINFODEBUGrXevalr lenrRrZrprar`logitsndimrlr functionalsoftmaxextendr;rPstacktensornpasarraycpudetachfloatnumpy)r]rrrr`rrrinput_was_singular pred_scores start_indexbatchfeaturesmodel_predictionsrscores r)rzCrossEncoder.predicts)b# Ie}=*YWX\[^B_" I!%   $((*gll:if>V>V>X\c\i\i>i   $  " "=e " D  !!S^Zi]nYno 'KkK*,DEE~~# &H KK )) * * HX H4 H ''(9(@(@AFq,,44V4C   v & ' ;; ! !Q &1<=58=K= ;#kk+6 #ll2djj6G6GH **Xc%dueiik&8&8&:&@&@&B&H&H&J%deK %a.K>&es $ I'AI,c D|jdk7r td|D cgc]} || g} } |j| ||||| | } g}t| D]5\}}|j ||d|s|dj d||i7t |dd }|d |Scc} w) a Performs ranking with the CrossEncoder on the given query and documents. Returns a sorted list with the document indices and scores. Args: query (str): A single query. documents (List[str]): A list of documents. top_k (Optional[int], optional): Return the top-k documents. If None, all documents are returned. Defaults to None. return_documents (bool, optional): If True, also returns the documents. If False, only returns the indices and scores. Defaults to False. batch_size (int, optional): Batch size for encoding. Defaults to 32. show_progress_bar (bool, optional): Output progress bar. Defaults to None. activation_fn ([type], optional): Activation function applied on the logits output of the CrossEncoder. If None, nn.Sigmoid() will be used if num_labels=1, else nn.Identity. Defaults to None. convert_to_numpy (bool, optional): Convert the output to a numpy matrix. Defaults to True. apply_softmax (bool, optional): If there are more than 2 dimensions and apply_softmax=True, applies softmax on the logits output. Defaults to False. convert_to_tensor (bool, optional): Convert the output to a tensor. Defaults to False. Returns: List[Dict[Literal["corpus_id", "score", "text"], Union[int, float, str]]]: A sorted list with the "corpus_id", "score", and optionally "text" of the documents. Example: :: from sentence_transformers import CrossEncoder model = CrossEncoder("cross-encoder/ms-marco-MiniLM-L6-v2") query = "Who wrote 'To Kill a Mockingbird'?" documents = [ "'To Kill a Mockingbird' is a novel by Harper Lee published in 1960. It was immediately successful, winning the Pulitzer Prize, and has become a classic of modern American literature.", "The novel 'Moby-Dick' was written by Herman Melville and first published in 1851. It is considered a masterpiece of American literature and deals with complex themes of obsession, revenge, and the conflict between good and evil.", "Harper Lee, an American novelist widely known for her novel 'To Kill a Mockingbird', was born in 1926 in Monroeville, Alabama. She received the Pulitzer Prize for Fiction in 1961.", "Jane Austen was an English novelist known primarily for her six major novels, which interpret, critique and comment upon the British landed gentry at the end of the 18th century.", "The 'Harry Potter' series, which consists of seven fantasy novels written by British author J.K. Rowling, is among the most popular and critically acclaimed books of the modern era.", "'The Great Gatsby', a novel written by American author F. Scott Fitzgerald, was published in 1925. The story is set in the Jazz Age and follows the life of millionaire Jay Gatsby and his pursuit of Daisy Buchanan." ] model.rank(query, documents, return_documents=True) :: [{'corpus_id': 0, 'score': 10.67858, 'text': "'To Kill a Mockingbird' is a novel by Harper Lee published in 1960. It was immediately successful, winning the Pulitzer Prize, and has become a classic of modern American literature."}, {'corpus_id': 2, 'score': 9.761677, 'text': "Harper Lee, an American novelist widely known for her novel 'To Kill a Mockingbird', was born in 1926 in Monroeville, Alabama. She received the Pulitzer Prize for Fiction in 1961."}, {'corpus_id': 1, 'score': -3.3099542, 'text': "The novel 'Moby-Dick' was written by Herman Melville and first published in 1851. It is considered a masterpiece of American literature and deals with complex themes of obsession, revenge, and the conflict between good and evil."}, {'corpus_id': 5, 'score': -4.8989105, 'text': "'The Great Gatsby', a novel written by American author F. Scott Fitzgerald, was published in 1925. The story is set in the Jazz Age and follows the life of millionaire Jay Gatsby and his pursuit of Daisy Buchanan."}, {'corpus_id': 4, 'score': -5.082967, 'text': "The 'Harry Potter' series, which consists of seven fantasy novels written by British author J.K. Rowling, is among the most popular and critically acclaimed books of the modern era."}] r:z|CrossEncoder.rank() only works for models with num_labels=1. Consider using CrossEncoder.predict() with input pairs instead.)rrrr`rrr) corpus_idrtextc |dS)Nrr.)xs r)z#CrossEncoder.rank..Bs ' r+T)rreverseN)rPrqr enumerateappendupdatesorted)r]query documentstop_kreturn_documentsrrr`rrrdocquery_doc_pairsscoresresultsirs r)rankzCrossEncoder.ranksH ??a R 4==CE3<==%!/''-/ !&) ;HAu NNU; < ""FIaL#9: ; &:DIv%>s Bsafe_serializationc |ytjd||jdt|jj |fd|i||j j |fi||j|y)zW Saves the model and tokenizer to path; identical to `save_pretrained` NzSave model to r r)rKrYrwrrpsave_pretrainedrR_create_model_cardr]pathrr&s r)savezCrossEncoder.saveEsv <  nTF+, i5" ""4Y Create an automatic model and stores it in the specified path. If no training was done and the loaded model was a CrossEncoder model already, then its model card is reused. Args: path (str): The path where the model card will be stored. Returns: None generated_from_trainerz-model = CrossEncoder("cross_encoder_model_id"zmodel = CrossEncoder(""z#Error while generating model card: zConsider opening an issue on https://github.com/UKPLab/sentence-transformers/issues with this traceback. Skipping model card creation.Nr?wrArB)rGrFtagsmodel_idreplacerrWrKerror traceback format_excrUr#rjoinwrite)r]r model_cardfOuts r)rzCrossEncoder._create_model_cardas  %=TEYEYE^E^%^..J##,,'//C,T-A-A-J-J,K1M  06 "'',,t[13 H #D JJz " # #  :9;O;O;Q:R44    # #s3 C*C?4C<;C<?D)r7privatercommit_messager"r6 create_prrc t|} | j||d|xs|} | j}|jj || |jj | || j ||d|d}d} tj5} |j| || j|| || || }ddd|r jSjS#1swY#xYw) a Upload the CrossEncoder model to the Hugging Face Hub. Example: :: from sentence_transformers import CrossEncoder model = CrossEncoder("cross-encoder/ms-marco-MiniLM-L6-v2") model.push_to_hub("username/my-crossencoder-model") # => "https://huggingface.co/username/my-crossencoder-model" Args: repo_id (str): The name of the repository on the Hugging Face Hub, e.g. "username/repo_name", "organization/repo_name" or just "repo_name". token (str, optional): The authentication token to use for the Hugging Face Hub API. If not provided, will use the token stored via the Hugging Face CLI. private (bool, optional): Whether to create a private repository. If not specified, the repository will be public. safe_serialization (bool, optional): Whether or not to convert the model weights in safetensors format for safer serialization. Defaults to True. commit_message (str, optional): The commit message to use for the push. Defaults to "Add new CrossEncoder model". exist_ok (bool, optional): If True, do not raise an error if the repository already exists. Ignored if ``create_pr=True``. Defaults to False. revision (str, optional): The git branch to commit to. Defaults to the head of the 'main' branch. create_pr (bool, optional): Whether to create a Pull Request with the upload or directly commit. Defaults to False. tags (list[str], optional): A list of tags to add to the model card. Defaults to None. Returns: str: URL of the commit or pull request (if create_pr=True) )r7N)repo_idr repo_typer"T)rbranchr"zAdd new CrossEncoder modelr)r folder_pathrcommit_descriptionr6r) r create_reporrF set_model_idadd_tags create_branchtempfileTemporaryDirectoryr upload_folderpr_url commit_url)r]rr7rrrr"r6rrapirepo_urlrtmp_dir folder_urls r) push_to_hubzCrossEncoder.push_to_hubsX% ??* #  "" ))'2    ) )$ /     gh  N  !9N  ( ( * g  #5 ! **#-#5!# +J  $$ $$$$!  s +C''C0c|jS)a Property to get the underlying transformers PreTrainedModel instance. Returns: PreTrainedModel or None: The underlying transformers model or None if not found. Example: :: from sentence_transformers import CrossEncoder model = CrossEncoder("cross-encoder/ms-marco-MiniLM-L6-v2") # You can now access the underlying transformers model transformers_model = model.transformers_model print(type(transformers_model)) # => rrss r)transformers_modelzCrossEncoder.transformers_models,zzr+cDtjd|jS)NzY`CrossEncoder._target_device` has been removed, please use `CrossEncoder.device` instead.)rKrLrarss r)_target_devicezCrossEncoder._target_devices g {{r+c&|j|yr)rZ)r]ras r)r zCrossEncoder._target_devices  r+c.|jjSr)rprarss r)razCrossEncoder.devicerr+c8|jj|Sr)rpgradient_checkpointing_enable)r]gradient_checkpointing_kwargss r)rz*CrossEncoder.gradient_checkpointing_enableszz778UVVr+)NNNNNFNFNNNNNrl) r^rrP int | Noner_rr`Callable | Nonera str | Noner@rr5boolr6rr3rr7zbool | str | Nonerb dict | NonercrrdrrFz CrossEncoderModelCardData | Noner<$Literal['torch', 'onnx', 'openvino']r,r-)r^rr;rr<rr,r-)r,r)T)r`rrxrr,r-)r,r)rrr,r-)r,r)r,int)rrr,r-)......)rztuple[str, str] | list[str]rrr bool | Noner`rrrrLiteral[False]rrr, torch.Tensor)....TF)rElist[tuple[str, str]] | list[list[str]] | tuple[str, str] | list[str]rrrrr`rrrr Literal[True]rrr,z np.ndarray)rrrrrrr`rrrrrrrr,r)rz'list[tuple[str, str]] | list[list[str]]rrrrr`rrrrrrrr,zlist[torch.Tensor]) NNFTF)rrrrrrr`rrrrrrrr,z.list[torch.Tensor] | np.ndarray | torch.Tensor)NFrNNFTF)rrrz list[str]rrrrrrrrr`rrrrrr,zDlist[dict[Literal['corpus_id', 'score', 'text'], int | float | str]])rrrrr,r-)rrr,r-)rrr7rrrrrrrr"rr6rrrrzlist[str] | Noner,r)r,zPreTrainedModel | None)r,z torch.devicer)razint | str | torch.device | Noner,r-)r,r-)"__name__ __module__ __qualname____doc__rrErQrtrXrvrwpropertyr;rPr_setterrrrrrrlinference_moderrrrrr r r rar __classcell__)rjs@r)r1r10sCJ'"&!%)-!#'"'#!&#'$((,%)<@8?!lSlSlS lS ' lS  lS!lS lSlSlS!lS"lS&lS#lS:lS 6!lS" #lS'lS\vv!v v  v<Q4 S!!&&//00 ."  "+),),%(+.,/ .  '  '  # ) *    ),),%(*.,1 X  '  '  # ( *    ),),%(!$+. X  '  '  #  )    ),),%(+.,/ !: ! !' ! ' ! # !) !* !  ! !U.)-)-%*!%"'\X\\' \ ' \ # \\ \ 8\/\|/ !!&)-)-!%"'ZZZ Z  Z  Z'Z'ZZ Z NZ/Zx=A &HL P!#N!##'%)#!%N%N% N%  N% ! N%#N%N%N%N%N% N%`. !!WWr+r1)r'rr(rr,zCallable[..., None])9 __future__rrr#rrpathlibrtypingrrrrrrlhuggingface_hubr packagingr r tqdm.autonotebookr transformersr r rrrrtransformers.utilsrtyping_extensionsrr8rsentence_transformers.backendrr-sentence_transformers.cross_encoder.fit_mixinr.sentence_transformers.cross_encoder.model_cardrr(sentence_transformers.cross_encoder.utilrrsentence_transformers.utilrrrr getLoggerr rKr/Moduler1r.r+r)r8s" .. !$.(-NBied   8 $M W299nhM Wr+