Ig-ddlmZmZmZmZmZmZmZmZddl Z ddl m Z ddl mZddlmZddlmZddlmZddlmZdd lmZGd d eZy) )AnyDictIterableListOptionalSequenceTypeUnionN) OnnxProvider)CLIPOnnxEmbedding)E5OnnxEmbedding)PooledNormalizedEmbedding)PooledEmbedding)OnnxTextEmbedding)TextEmbeddingBasec eZdZUeeeeegZe e e e d<e de eeeffdZ ddedeedeedeeeded ee ed effd Z dd eeeefd edeedeej6fdZxZS) TextEmbeddingEMBEDDINGS_REGISTRYreturncjg}|jD]!}|j|j#|S)aa Lists the supported models. Returns: List[Dict[str, Any]]: A list of dictionaries containing the model information. Example: ``` [ { "model": "intfloat/multilingual-e5-large", "dim": 1024, "description": "Multilingual model, e5-large. Recommend using this model for non-English languages", "license": "mit", "size_in_GB": 2.24, "sources": { "gcp": "https://storage.googleapis.com/qdrant-fastembed/fast-multilingual-e5-large.tar.gz", "hf": "qdrant/multilingual-e5-large-onnx", } } ] ``` )rextendlist_supported_models)clsresult embeddings Z/var/www/html/answerous/venv/lib/python3.12/site-packages/fastembed/text/text_embedding.pyrz#TextEmbedding.list_supported_modelss8200 =I MM)99; < = model_name cache_dirthreads providerscuda device_ids lazy_loadc t |||fi||jD]=} | j} t fd| Ds(| d||||||d||_yt dd)Nc3fK|](}j|djk(*yw)modelN)lower).0r'rs r z)TextEmbedding.__init__..Cs,^E:##%w)=)=)??^s.1)rrr r!r"r#r$zModel zr is not supported in TextEmbedding.Please check the supported models using `TextEmbedding.list_supported_models()`)super__init__rranyr' ValueError) selfrrr r!r"r#r$kwargsEMBEDDING_MODEL_TYPEsupported_models __class__s ` rr-zTextEmbedding.__init__5s YB6B$($<$<  3IIK ^M]^^1 )'#')'    ZL!^ ^  r documents batch_sizeparallelc+^K|jj|||fi|Ed{y7w)a Encode a list of documents into list of embeddings. We use mean pooling with attention so that the model can handle variable-length inputs. Args: documents: Iterator of documents or single document to embed batch_size: Batch size for encoding -- higher values will use more memory, but be faster parallel: If > 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 N)r'embed)r0r5r6r7r1s rr9zTextEmbedding.embedUs+,$4::##Iz8NvNNNs #-+-)zBAAI/bge-small-en-v1.5NNNFNF)N)__name__ __module__ __qualname__rr r rrrrr r__annotations__ classmethodrstrrrrintrr boolr-r rnpndarrayr9 __classcell__)r4s@rrrs#! :d#456d4S>&:>3#'!%6:*.  C= #  H\23    T#Y'  F"& Ohsm+,OO3- O "**  Orr)typingrrrrrrr r numpyrCfastembed.commonr fastembed.text.clip_embeddingr fastembed.text.e5_onnx_embeddingr *fastembed.text.pooled_normalized_embeddingrfastembed.text.pooled_embeddingrfastembed.text.onnx_embeddingr"fastembed.text.text_embedding_baserrr+rrrOs7MMM);<P;;@]O%]Or