o wiQ@sddlmZmZddlmZddlmZmZmZddl m Z m Z m Z m Z ddlmZmZmZddlmZer;ddlZGd d d e d d Zd ZddedDddedDZddZGddde ZdgZdS))OptionalUnion) BatchFeature) ImageInputis_valid_imagemake_flat_list_of_images)MultiModalDataProcessingKwargsProcessorMixinUnpack) AddedTokenPreTokenizedInput TextInput)is_torch_availableNc@s&eZdZddidddddidZd S) ColPaliProcessorKwargspaddinglongestchannels_firstT) data_formatdo_convert_rgbreturn_tensorspt) text_kwargs images_kwargs common_kwargsN)__name__ __module__ __qualname__ _defaultsr r k/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/transformers/models/colpali/processing_colpali.pyr$s rF)totalzcCg|] }d|ddqS)z4>r .0ir r r! 2r(icCr#)z3r$r r%r r r!r(2r)cCs|||||dS)aZ Builds a string from the input prompt and image tokens. For example, for the call: build_string_from_input( prompt="Prefix str" bos_token="", image_seq_len=3, image_token="", ) The output will be: "Initial str" Args: prompt (`list[Union[str, ImageInput]]`): The input prompt. bos_token (`str`): The beginning of sentence token. image_seq_len (`int`): The length of the image sequence. image_token (`str`): The image token. num_images (`int`): Number of images in the prompt.  r prompt bos_token image_seq_len image_token num_imagesr r r!build_string_from_input5sr2c sLeZdZdZddgZdZdZ     d-d ed effd d Z    d.d e de e e e e e e fdeedefddZd/ddZddZddZeddZedefddZ d/d e deedefddZde e e e fdeedefdd Z !  "d0d#e d$e d$fd%e d$e d$fd&ed'ed(d)e d*efdd$f d+d,ZZS)1ColPaliProcessora Constructs a ColPali processor which wraps a PaliGemmaProcessor and special methods to process images and queries, as well as to compute the late-interaction retrieval score. [`ColPaliProcessor`] offers all the functionalities of [`PaliGemmaProcessor`]. See the [`~PaliGemmaProcessor.__call__`] for more information. Args: image_processor ([`SiglipImageProcessor`], *optional*): The image processor is a required input. tokenizer ([`LlamaTokenizerFast`], *optional*): The tokenizer is a required input. chat_template (`str`, *optional*): A Jinja template which will be used to convert lists of messages in a chat into a tokenizable string. visual_prompt_prefix (`str`, *optional*, defaults to `"Describe the image."`): A string that gets tokenized and prepended to the image tokens. query_prefix (`str`, *optional*, defaults to `"Question: "`): A prefix to be used for the query. image_processor tokenizer)SiglipImageProcessorSiglipImageProcessorFast)GemmaTokenizerGemmaTokenizerFastNDescribe the image. Question: visual_prompt_prefix query_prefixcstj|||d|durtd|durtdt|ds"td|j|_t|dsFttddd }d |gi}|||t|_ t|_ n|j |_ |j |_ | t d|_ d|_||_||_dS) N) chat_templatez)You need to specify an `image_processor`.z"You need to specify a `tokenizer`.image_seq_lengthz;Image processor is missing an `image_seq_length` attribute.r0FT) normalizedspecialadditional_special_tokens)super__init__ ValueErrorhasattrr?r IMAGE_TOKENadd_special_tokensconvert_tokens_to_idsimage_token_idr0 add_tokens EXTRA_TOKENS add_bos_token add_eos_tokenr<r=)selfr4r5r>r<r=r0 tokens_to_add __class__r r!rDds*       zColPaliProcessor.__init__imagestextkwargsreturnc sfjtfdjji|}|ddd}|durdnd}|dur)|dur)td|dur5|dur5td|durt|rA|g}n$t|trMt|d rMnt|trat|d trat|d d setd j gt |} d d |D}fd d t | |D} t |}j |fi|dd} |ddddur|ddj7<j| fddi|d} i| d| i} |r| d| dd kd}| d|it| dS|dur1t|tr|g}nt|trt|d tstd|durjd}g}|D]}jjj||d}||q|ddd|dd<j|fddi|d}|SdS)a Main method to prepare for the model either (1) one or several texts, either (2) one or several image(s). This method is a custom wrapper around the PaliGemmaProcessor's [`~PaliGemmaProcessor.__call__`] method adapted for the ColPali model. It cannot process both text and images at the same time. When preparing the text(s), this method forwards the `text` and `kwargs` arguments to LlamaTokenizerFast's [`~LlamaTokenizerFast.__call__`]. When preparing the image(s), this method forwards the `images` and `kwargs` arguments to SiglipImageProcessor's [`~SiglipImageProcessor.__call__`]. Please refer to the docstring of the above two methods for more information. Args: images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `list[PIL.Image.Image]`, `list[np.ndarray]`, `list[torch.Tensor]`): The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch tensor. In case of a NumPy array/PyTorch tensor, each image should be of shape (C, H, W), where C is a number of channels, H and W are image height and width. text (`str`, `list[str]`, `list[list[str]]`): The sequence or batch of sequences to be encoded. Each sequence can be a string or a list of strings (pretokenized string). If the sequences are provided as list of strings (pretokenized), you must set `is_split_into_words=True` (to lift the ambiguity with a batch of sequences). return_tensors (`str` or [`~utils.TensorType`], *optional*): If set, will return tensors of a particular framework. Acceptable values are: - `'tf'`: Return TensorFlow `tf.constant` objects. - `'pt'`: Return PyTorch `torch.Tensor` objects. - `'np'`: Return NumPy `np.ndarray` objects. - `'jax'`: Return JAX `jnp.ndarray` objects. Returns: [`BatchFeature`]: A [`BatchFeature`] with the following fields: - **input_ids** -- List of token ids to be fed to a model. - **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when `return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not `None`). - **pixel_values** -- Pixel values to be fed to a model. Returned when `images` is not `None`. tokenizer_init_kwargsrsuffixNTFz&Either text or images must be providedz5Only one of text or images can be processed at a timerzAimages must be an image, list of images or list of list of imagescSsg|]}|dqS)RGB)convert)r&imager r r!r(sz-ColPaliProcessor.__call__..c s:g|]\}}t|jjjtt|trt|nddqS)r,)r2r5r.r?rG isinstancelistlen)r&r- image_listrOr r!r(sr pixel_values max_lengthreturn_token_type_ids input_idstoken_type_idsilabels)dataz*Text must be a string or a list of strings r+2) _merge_kwargsrr5 init_kwargspoprErr]r^r<r_ziprr4getr? masked_fillupdaterstrquery_augmentation_tokenr.r=append)rOrSrTaudiovideosrU output_kwargsrXrd texts_doc input_stringsrbinputs return_datarg texts_queryquery batch_queryr rar!__call__s|-(      zColPaliProcessor.__call__cKsHi}|dur|jgt|}dgt|}|||dtdi|S)aB Computes the number of placeholder tokens needed for multimodal inputs with the given sizes. Args: image_sizes (list[list[str]], *optional*): The input sizes formatted as (height, width) per each image. Returns: dict[str, list[int]]: A dictionary mapping each modality ("image", "video", "audio") to a list containing the number of placeholder tokens required. If the model doesn't accept a certain modality or no input sizes are provided, the dict value is set to an empty list. Nr\)num_image_tokensnum_image_patchesr )r?r_rqr )rO image_sizesrU vision_datarrr r r!_get_num_multimodal_tokenss  z+ColPaliProcessor._get_num_multimodal_tokenscO|jj|i|S)z This method forwards all its arguments to GemmaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please refer to the docstring of this method for more information. )r5 batch_decoderOargsrUr r r!rzColPaliProcessor.batch_decodecOr)z This method forwards all its arguments to GemmaTokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to the docstring of this method for more information. )r5decoderr r r!rrzColPaliProcessor.decodecCs"|jj}|jj}tt||SN)r5model_input_namesr4r^dictfromkeys)rOtokenizer_input_namesimage_processor_input_namesr r r!r$sz"ColPaliProcessor.model_input_namescCs|jjS)z Return the query augmentation token. Query augmentation buffers are used as reasoning buffers during inference. )r5 pad_tokenrar r r!rs*sz)ColPaliProcessor.query_augmentation_tokencK|jdd|i|S)a Prepare for the model one or several image(s). This method is a wrapper around the `__call__` method of the ColPaliProcessor's [`ColPaliProcessor.__call__`]. This method forwards the `images` and `kwargs` arguments to the image processor. Args: images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `list[PIL.Image.Image]`, `list[np.ndarray]`, `list[torch.Tensor]`): The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch tensor. In case of a NumPy array/PyTorch tensor, each image should be of shape (C, H, W), where C is a number of channels, H and W are image height and width. return_tensors (`str` or [`~utils.TensorType`], *optional*): If set, will return tensors of a particular framework. Acceptable values are: - `'tf'`: Return TensorFlow `tf.constant` objects. - `'pt'`: Return PyTorch `torch.Tensor` objects. - `'np'`: Return NumPy `np.ndarray` objects. - `'jax'`: Return JAX `jnp.ndarray` objects. Returns: [`BatchFeature`]: A [`BatchFeature`] with the following fields: - **input_ids** -- List of token ids to be fed to a model. - **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when `return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not `None`). - **pixel_values** -- Pixel values to be fed to a model. Returned when `images` is not `None`. rSNr r)rOrSrUr r r!process_images3s!zColPaliProcessor.process_imagescKr)a Prepare for the model one or several texts. This method is a wrapper around the `__call__` method of the ColPaliProcessor's [`ColPaliProcessor.__call__`]. This method forwards the `text` and `kwargs` arguments to the tokenizer. Args: text (`str`, `list[str]`, `list[list[str]]`): The sequence or batch of sequences to be encoded. Each sequence can be a string or a list of strings (pretokenized string). If the sequences are provided as list of strings (pretokenized), you must set `is_split_into_words=True` (to lift the ambiguity with a batch of sequences). return_tensors (`str` or [`~utils.TensorType`], *optional*): If set, will return tensors of a particular framework. Acceptable values are: - `'tf'`: Return TensorFlow `tf.constant` objects. - `'pt'`: Return PyTorch `torch.Tensor` objects. - `'np'`: Return NumPy `np.ndarray` objects. - `'jax'`: Return JAX `jnp.ndarray` objects. Returns: [`BatchFeature`]: A [`BatchFeature`] with the following fields: - **input_ids** -- List of token ids to be fed to a model. - **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when `return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not `None`). rTNr r)rOrTrUr r r!process_queriesVs z ColPaliProcessor.process_queriesr*cpuquery_embeddingsz torch.Tensorpassage_embeddings batch_size output_dtypez torch.dtype output_devicez torch.devicec Cs@t|dkr tdt|dkrtd|dj|djkr"td|dj|djkr0td|dur9|dj}g}tdt||D]U}g}tjjjj ||||ddd} tdt||D]'} tjjjj || | |ddd} | t d | | j d d dj d d q`| tj|d d ||qCtj|dd S)aZ Compute the late-interaction/MaxSim score (ColBERT-like) for the given multi-vector query embeddings (`qs`) and passage embeddings (`ps`). For ColPali, a passage is the image of a document page. Because the embedding tensors are multi-vector and can thus have different shapes, they should be fed as: (1) a list of tensors, where the i-th tensor is of shape (sequence_length_i, embedding_dim) (2) a single tensor of shape (n_passages, max_sequence_length, embedding_dim) -> usually obtained by padding the list of tensors. Args: query_embeddings (`Union[torch.Tensor, list[torch.Tensor]`): Query embeddings. passage_embeddings (`Union[torch.Tensor, list[torch.Tensor]`): Passage embeddings. batch_size (`int`, *optional*, defaults to 128): Batch size for computing scores. output_dtype (`torch.dtype`, *optional*, defaults to `torch.float32`): The dtype of the output tensor. If `None`, the dtype of the input embeddings is used. output_device (`torch.device` or `str`, *optional*, defaults to "cpu"): The device of the output tensor. Returns: `torch.Tensor`: A tensor of shape `(n_queries, n_passages)` containing the scores. The score tensor is saved on the "cpu" device. rzNo queries providedzNo passages providedz/Queries and passages must be on the same devicez-Queries and passages must have the same dtypeNT) batch_first padding_valuez bnd,csd->bcnsr)dimr\)r_rEdevicedtyperangetorchnnutilsrnn pad_sequencerteinsummaxsumcatto) rOrrrrrscoresr' batch_scores batch_queriesjbatch_passagesr r r!score_retrievalxs2      "z ColPaliProcessor.score_retrieval)NNNr:r;)NNNNr)r*Nr)rrr__doc__ attributesimage_processor_classtokenizer_classrrrDrrrrr^r rrrrrrpropertyrrsrrintrr __classcell__r r rQr!r3Ks$ }   # & r3)typingrrfeature_extraction_utilsr image_utilsrrrprocessing_utilsr r r r tokenization_utils_baser rrrrrrrGrrLr2r3__all__r r r r!s    $ p