U 9%eX@sddlmZmZddlmZmZmZmZmZm Z ddl m Z m Z er^ddl mZddlmZerpddlmZerdd lZdd lmZeeZee Gd d d e Zd S) )ListUnion)add_end_docstringsis_tf_availableis_torch_availableis_vision_availableloggingrequires_backends)PIPELINE_INIT_ARGSPipeline)Image) load_image)'TF_MODEL_FOR_VISION_2_SEQ_MAPPING_NAMESN)$MODEL_FOR_VISION_2_SEQ_MAPPING_NAMEScsleZdZdZfddZdddZeeeededfdfd d Z dd d Z dd dZ ddZ Z S)ImageToTextPipelinea Image To Text pipeline using a `AutoModelForVision2Seq`. This pipeline predicts a caption for a given image. Example: ```python >>> from transformers import pipeline >>> captioner = pipeline(model="ydshieh/vit-gpt2-coco-en") >>> captioner("https://huggingface.co/datasets/Narsil/image_dummy/raw/main/parrots.png") [{'generated_text': 'two birds are standing next to each other '}] ``` Learn more about the basics of using a pipeline in the [pipeline tutorial](../pipeline_tutorial) This image to text pipeline can currently be loaded from pipeline() using the following task identifier: "image-to-text". See the list of available models on [huggingface.co/models](https://huggingface.co/models?pipeline_tag=image-to-text). cs4tj||t|d||jdkr*tntdS)NZvisiontf)super__init__r Zcheck_model_type frameworkrr)selfargskwargs __class__c/var/www/html/Darija-Ai-API/env/lib/python3.8/site-packages/transformers/pipelines/image_to_text.pyr6s  zImageToTextPipeline.__init__NcCszi}i}|dk r||d<|dk r(||d<|dk r8||d<|dk rpd|krPi|d<d|dkrdtd||dd<||ifS)Nprompttimeoutgenerate_kwargsmax_new_tokenszp'max_new_tokens' is defined twice, once in 'generate_kwargs' and once as a direct parameter, please use only one) ValueError)rr!r rrZforward_kwargsZpreprocess_paramsrrr_sanitize_parameters=s"  z(ImageToTextPipeline._sanitize_parametersz Image.Image)imagesc stj|f|S)a Assign labels to the image(s) passed as inputs. Args: images (`str`, `List[str]`, `PIL.Image` or `List[PIL.Image]`): The pipeline handles three types of images: - A string containing a HTTP(s) link pointing to an image - A string containing a local path to an image - An image loaded in PIL directly The pipeline accepts either a single image or a batch of images. max_new_tokens (`int`, *optional*): The amount of maximum tokens to generate. By default it will use `generate` default. generate_kwargs (`Dict`, *optional*): Pass it to send all of these arguments directly to `generate` allowing full control of this function. timeout (`float`, *optional*, defaults to None): The maximum time in seconds to wait for fetching images from the web. If None, no timeout is set and the call may block forever. Return: A list or a list of list of `dict`: Each result comes as a dictionary with the following key: - **generated_text** (`str`) -- The generated text. )r__call__)rr$rrrrr%SszImageToTextPipeline.__call__cCs*t||d}|dk rt|ts2tdt|d|jjj}|dkr|j||j d}|j |ddj }|j j g|}t |d}|d |in`|d kr|j|||j d }nD|d kr|j||j d}|j ||j d }||ntd|dn|j||j d}|jjjdkr&|dkr&d|d <|S)N)rz&Received an invalid text input, got - zy - but expected a single string. Note also that one single text can be provided for conditional image to text generation.git)r$return_tensorsF)textZadd_special_tokensr input_idsZ pix2struct)r$Z header_textr'zvision-encoder-decoder)r'z Model type z- does not support conditional text generation)r isinstancestrr"typemodelconfig model_typeZimage_processorr tokenizerr)Z cls_token_idtorchZtensorZ unsqueezeupdate)rimagerrr/ model_inputsr)Z text_inputsrrr preprocessqs0    zImageToTextPipeline.preprocesscCsfd|kr4t|dtr4tdd|dDr4d|d<|dkr@i}||jj}|jj|f||}|S)Nr)css|]}|dkVqdS)Nr).0xrrr sz/ImageToTextPipeline._forward..)r*listallpopr-Zmain_input_namegenerate)rr4r inputs model_outputsrrr_forwards zImageToTextPipeline._forwardcCs0g}|D]"}d|jj|ddi}||q|S)NZgenerated_textT)Zskip_special_tokens)r0decodeappend)rr>recordsZ output_idsrecordrrr postprocesss zImageToTextPipeline.postprocess)NNNN)NN)N)__name__ __module__ __qualname____doc__rr#rr+rr%r5r?rD __classcell__rrrrrs  & ' r)typingrrutilsrrrrr r baser r ZPILrZ image_utilsrZmodels.auto.modeling_tf_autorr1Zmodels.auto.modeling_autorZ get_loggerrEloggerrrrrrs