qiddlmZddlmZddlmZmZmZddlm Z m Z ddl m Z m Z e jeZGdded Ze Gd d eZd gZy ) ) BatchFeature) ImageInput)ProcessingKwargsProcessorMixinUnpack)PreTokenizedInput TextInput)auto_docstringloggingceZdZddiidZy)Ovis2ProcessorKwargspaddingF) text_kwargs image_kwargsN)__name__ __module__ __qualname__ _defaults\/opt/pipecat/venv/lib/python3.12/site-packages/transformers/models/ovis2/processing_ovis2.pyr r s u  Irr F)totalc eZdZ d fd Ze ddedzdeezeezeezde e de fdZ deedeee fd Zd Zd Zed ZxZS)Ovis2ProcessorNc ||_t|dr |jn||_t|ddr |jn|j |j|_t |||fd|i|y)a image_token (`str`, *optional*, defaults to `""`): Special token used to denote image location. image_seq_length (`int`, *optional*, defaults to 256): The number of image tokens to be used for each image in the input. image_tokenimage_token_idN chat_template)image_seq_lengthhasattrrgetattrrconvert_tokens_to_idssuper__init__)selfimage_processor tokenizerrrrkwargs __class__s rr$zOvis2Processor.__init__%sy!14;I}4U900[fy"2D9  $ $001A1AB  )[=[TZ[rimagestextr(returnc |jtfd|jji|}t |t r|g}n.t |t st |dt s tdi}|G|j|fi|d}|jdj}|j||}|j|fi|d}ti||S)a Returns: [`BatchFeature`]: A [`BatchFeature`] with the following fields: - **input_ids** -- List of token ids to be fed to a model. Returned when `text` is not `None`. - **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when `return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not `None`). - **pixel_values** -- Pixel values to be fed to a model. Returned when `images` is not `None`. - **image_sizes** -- Size of each image that will be used to unpad an image. Returned when `images` is not `None`. tokenizer_init_kwargszAInvalid input text. Please provide a string, or a list of strings images_kwargsgridsr)data) _merge_kwargsr r' init_kwargs isinstancestrlist TypeErrorr&poptolist_expand_image_tokensr)r%r*r+r( output_kwargs image_inputs image_grids text_inputss r__call__zOvis2Processor.__call__=s&+**  "&.."<"<  dC 6DD$' 47C0H_` `  /4//Y-:XYL&**73::z z z z z z )rrangereplaceappend) r%r+r1processed_text grid_indexsamplegridrowcol placeholderrcs rr;z#Ovis2Processor._expand_image_tokenses   *Fv%Z(7DGS +L4;P;P,P+QQ[\ 9q="3Z7!&s;A'lT=R=R.R-STK 37{ +{ : ;sQw;';6K 7{*  ;Ba v%  ! !& )# *$rc:|jj|i|S)z This method forwards all its arguments to Qwen2TokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please refer to the docstring of this method for more information. )r' batch_decoder%argsr(s rrQzOvis2Processor.batch_decodes +t~~**D;F;;rc:|jj|i|S)z This method forwards all its arguments to Qwen2TokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to the docstring of this method for more information. )r'decoderRs rrUzOvis2Processor.decodes %t~~$$d5f55rc|jj}|jj}t|t|zS)N)r'model_input_namesr&r7)r%tokenizer_input_namesimage_processor_input_namess rrWz Ovis2Processor.model_input_namess; $ @ @&*&:&:&L&L#)*T2M-NNNr)NNNrB)NN)rrrr$r rr rr7rr rr@intr;rQrUpropertyrW __classcell__)r)s@rrr#s \0%)Z^%BT!%B++d9o=EV@WW%B-. %B  %B%BN9oDI6<6OOrrN)feature_extraction_utilsr image_utilsrprocessing_utilsrrrtokenization_utils_baserr utilsr r get_loggerrloggerr r__all__rrrrfsh 5%HHC,   H %+5nO^nOnOb  r