^Nj (UddlmZmZmZmZddlZddlmZddl m Z ddl m Z ddl mZddlmZmZddlmZmZdd lmZmZed d d d deddedd d d deddgZeeed<GddeeeZGddeeZy))AnyIterableSequenceTypeN) OnnxProvider)OnnxOutputContext)Device)define_cache_dir)SparseEmbeddingSparseTextEmbeddingBase) OnnxTextModelTextEmbeddingWorker)SparseModelDescription ModelSourcezprithivida/Splade_PP_en_v1i:wz9Independent Implementation of SPLADE++ Model for English.z apache-2.0g/$?zQdrant/Splade_PP_en_v1)hfz model.onnx)model vocab_size descriptionlicense size_in_GBsources model_filezprithvida/Splade_PP_en_v1supported_splade_modelscfeZdZdededeefdZ ddeeezde dede fdZ e de e fdZd d d ej d d d d fd ed ed zd e d zdeed zdeezde e d zdede d zded zdeffd ZddZ ddeeezde de d zdedeef dZe deeefdZxZS)SpladePPoutputkwargsreturnc+xK|j tdtjdtj|j dz}|tj |jdz}tj|d}|D])}|jd}||}t||+yw)Nzs '&r:NF model_name cache_dirthreads providerscuda device_ids lazy_load device_idspecific_model_pathc  t ||||fi| ||_||_|j | |_||_||_d|_|||_n |j |j d|_|j||_ tt||_ | |_|j|j|j|j |j|_|js|j%yy)a Args: model_name (str): The name of the model to use. cache_dir (str, optional): The path to the cache directory. Can be set using the `FASTEMBED_CACHE_PATH` env variable. Defaults to `fastembed_cache` in the system's temp directory. threads (int, optional): The number of threads single onnxruntime session can use. Defaults to None. providers (Optional[Sequence[OnnxProvider]], optional): The list of onnxruntime providers to use. Mutually exclusive with the `cuda` and `device_ids` arguments. Defaults to None. cuda (Union[bool, Device], optional): Whether to use cuda for inference. Mutually exclusive with `providers` Defaults to Device. device_ids (Optional[list[int]], optional): The list of device ids to use for data parallel processing in workers. Should be used with `cuda` equals to `True`, `Device.AUTO` or `Device.CUDA`, mutually exclusive with `providers`. Defaults to None. lazy_load (bool, optional): Whether to load the model during class initialization or on demand. Should be set to True when using multiple-gpu and parallel encoding. Defaults to False. device_id (Optional[int], optional): The device id to use for loading the model in the worker process. specific_model_path (Optional[str], optional): The specific path to the onnx model dir if it should be imported from somewhere else Raises: ValueError: If the model_name is not in the format / e.g. BAAI/bge-base-en. Nr)local_files_onlyrG)super__init__rBrE_select_exposed_session_options_extra_session_optionsrDrCrF_get_model_descriptionmodel_descriptionstrr r@_specific_model_pathdownload_model_local_files_only _model_dirload_onnx_model) r.r?r@rArBrCrDrErFrGr __class__s r3rKzSpladePP.__init__GsF YB6B""&*&J&J6&R#% &*  &DN __ (!__Q/DN!%! 1, data-parallel encoding will be used, recommended for offline encoding of large datasets. If 0, use all available cores. If None, don't use data-parallel processing, use default onnxruntime threading instead. Returns: List of embeddings, one per document ) r?r@r[r6r\rBrCrDrIrGrYN) _embed_documentsr?rPr@rBrCrDrSrQrM)r.r[r6r\rs r3embedzSpladePP.embeds},)4((  $..)!nn!33 $ 9 9"&"="=    sA;B=B>BctS)N)SpladePPEmbeddingWorkerr<s r3_get_worker_classzSpladePP._get_worker_classs&&r:)i)rN)N)__name__ __module__ __qualname__rrrr r4rPintr9 classmethodlistrr>r AUTOrrboolrKrUr`rrrc __classcell__)rVs@r3rr%sB'B36B / "B(=AI8C=(I69ILOI I 't,B'C''!%"37$kk'+ $*.?#?#:?#t ?# L)D0 ?# Vm ?#I$?#?#:?#!4Z?#?#B  # # #&# # * #  # / " # J'$':?'K"L''r:rc$eZdZdedededefdZy)rbr?r@rrc td||dd|S)Nr )r?r@rAr^)r)r.r?r@rs r3init_embeddingz&SpladePPEmbeddingWorker.init_embeddings' !    r:N)rerfrgrPrrrpr^r:r3rbrbs$    PX r:rb)typingrrrrnumpyr'fastembed.commonrfastembed.common.onnx_modelrfastembed.common.typesr fastembed.common.utilsr &fastembed.sparse.sparse_embedding_baser r fastembed.text.onnx_text_modelr r"fastembed.common.model_descriptionrrrrj__annotations__rrbr^r:r3r{s00)9)3NR*O78)O78945,U'& o(FU'p 1/B r: