o ZhO3@sdZddlmZddlmZmZmZmZddlm Z ddl m Z m Z m Z ddlmZmZmZmZmZddlmZmZdd lmZerHdd lmZeeZd efd d ZddZGdddeddZ GdddeddZ!GdddeZ"dgZ#dS)z Processor class for IDEFICS2. ) accumulate) TYPE_CHECKINGListOptionalUnion) BatchFeature) ImageInputis_valid_image load_image) ImagesKwargsProcessingKwargsProcessorMixinUnpack!_validate_images_text_input_order) AddedToken TextInput)logging)PreTokenizedInputreturncCst|to |dS)Nhttp) isinstancestr startswith)valr_/var/www/auris/lib/python3.10/site-packages/transformers/models/idefics2/processing_idefics2.pyis_url*srcCst|pt|SN)rr )elemrrris_image_or_image_url.sr c@seZdZUeeed<dS)Idefics2ImagesKwargs image_seq_lenN)__name__ __module__ __qualname__rint__annotations__rrrrr!2s r!F)totalc@s(eZdZUeed<ddddidZdS)Idefics2ProcessorKwargs images_kwargsTF)add_special_tokenspaddingZis_split_into_words) text_kwargsr*N)r#r$r%r!r' _defaultsrrrrr)6s  r)c seZdZdZddgZddgZdZdZ ddede e ffd d Z d d Z    dde eeeeeefde edeeedfdeedefddZddZddZeddZZS)Idefics2Processora Constructs a IDEFICS2 processor which wraps a LLama tokenizer and IDEFICS2 image processor into a single processor. [`IdeficsProcessor`] offers all the functionalities of [`Idefics2ImageProcessor`] and [`LlamaTokenizerFast`]. See the docstring of [`~IdeficsProcessor.__call__`] and [`~IdeficsProcessor.decode`] for more information. Args: image_processor (`Idefics2ImageProcessor`): An instance of [`Idefics2ImageProcessor`]. The image processor is a required input. tokenizer (`PreTrainedTokenizerBase`, *optional*): An instance of [`PreTrainedTokenizerBase`]. This should correspond with the model's text model. The tokenizer is a required input. image_seq_len (`int`, *optional*, defaults to 64): The length of the image sequence i.e. the number of tokens per image in the input. This parameter is used to build the string from the input prompt and image tokens and should match the config.perceiver_config.resampler_n_latents value for the model used. chat_template (`str`, *optional*): A Jinja template which will be used to convert lists of messages in a chat into a tokenizable string. image_processor tokenizerr" chat_templateZIdefics2ImageProcessorZ AutoTokenizerN@c s|durtd|durtdt|dsFT) normalizedZspecialzZadditional_special_tokensz)r2) ValueErrorhasattrrcontentfake_image_tokenr4r+Zconvert_tokens_to_idsZimage_token_idZimage_boundary_tokenZend_of_utterance_tokenr"super__init__)selfr0r1r"r2kwargsZ tokens_to_add __class__rrr;\s"  zIdefics2Processor.__init__cCsTg}|D]#}g}|D]}t|r||q t|r!|t|q ||q|Sr)r appendrr )r<ZpromptsZ prompt_imagespromptimagesrrrr_extract_images_from_promptsus  z.Idefics2Processor._extract_images_from_promptsrBtextrr=rc s|dur dur tdt|\}|jtfd|jji|}|ddd}|dur.|n|j}|ddd}g} i} |durt|t rJ|g}nt|t sZt|dt sZtd |j } |j } | | || } |j jrv| d } |d 9}g}|D] }| || || | }|| | | }||qz|j|fi|d}|j||d gd | |duratrˆggnltt tfrtdr|dur t| tkrtd | dt| d| dtd dgt t| fddtt| Dn#gntt tfs.tdt tfs.tdds.tdddD}|durJ|| ksJtd| d|dddD|j fi|d}| |t| |dS)a Processes the input prompts and returns a BatchEncoding. Example: ```python >>> import requests >>> from transformers import Idefics2Processor >>> from transformers.image_utils import load_image >>> processor = Idefics2Processor.from_pretrained("HuggingFaceM4/idefics2-8b", image_seq_len=2) >>> processor.image_processor.do_image_splitting = False # Force as False to simplify the example >>> url1 = "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" >>> url2 = "https://cdn.britannica.com/59/94459-050-DBA42467/Skyline-Chicago.jpg" >>> image1, image2 = load_image(url1), load_image(url2) >>> images = [[image1], [image2]] >>> text = [ ... "In this image, we see", ... "bla bla bla", ... ] >>> outputs = processor(images=images, text=text, return_tensors="pt", padding=True) >>> input_ids = outputs.input_ids >>> input_tokens = processor.tokenizer.batch_decode(input_ids) >>> print(input_tokens) [' In this image, we see', ' bla bla bla'] ``` Args: images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`, *optional*): The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch tensor. If is of type `List[ImageInput]`, it's assumed that this is for a single prompt i.e. of batch size 1. text (`Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]]`, *optional*): 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). Wherever an image token, `` is encountered it is expanded to `` + `` * `image_seq_len` * `. return_tensors (`Union[str, TensorType]`, *optional*): If set, will return tensors of a particular framework. See [`PreTrainedTokenizerFast.__call__`] for more information. Nz+You must provide either `text` or `images`.Ztokenizer_init_kwargsr*r"r-return_tensorsrzAInvalid input text. Please provide a string, or a list of stringsimage)Z modalitieszThe total number of zP tokens in the prompts should be the same as the number of images passed. Found  z tokens and z images.cs$g|]}||dqS)r).0iZcumsum_images_in_textrBrr sz.Idefics2Processor.__call__..zdInvalid input images. Please provide a single image or a list of images or a list of list of images.cSg|]}t|qSr)lenrJsamplerrrrMz!The number of images in the text z and images z should be the same.cSsg|] }dd|DqS)cSrNr)r )rJZimrrrrMrRz9Idefics2Processor.__call__...rrPrrrrMs)Z tensor_type)r6rZ _merge_kwargsr)r1Z init_kwargspopr"rrlistr9r4r0Zdo_image_splittingr@countreplaceZ_check_special_mm_tokensupdater tuplesumrOrranger)r<rBrDZaudioZvideosr=Z output_kwargsr"rEZn_images_in_textZinputsr9r4Z image_strZprompt_stringsrQZ text_inputsZn_images_in_imagesZ image_inputsrrLr__call__s6            zIdefics2Processor.__call__cO|jj|i|S)z This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please refer to the docstring of this method for more information. )r1 batch_decoder<argsr=rrrr] zIdefics2Processor.batch_decodecOr\)z This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to the docstring of this method for more information. )r1decoder^rrrrar`zIdefics2Processor.decodecCs"|jj}|jj}tt||Sr)r1model_input_namesr0rTdictfromkeys)r<Ztokenizer_input_namesZimage_processor_input_namesrrrrbsz#Idefics2Processor.model_input_names)Nr3N)NNNN)r#r$r%__doc__ attributesZ valid_kwargsZimage_processor_classZtokenizer_classr&rrr;rCrr rrrr)rr[r]rapropertyrb __classcell__rrr>rr/Cs>  r/N)$re itertoolsrtypingrrrrZfeature_extraction_utilsrZ image_utilsr r r Zprocessing_utilsr r rrrZtokenization_utils_baserrutilsrrZ get_loggerr#loggerboolrr r!r)r/__all__rrrrs$       `