T^(h(dZddlZddlmZmZddlmZddlmZddl m Z m Z m Z ddl mZmZmZmZdd lmZd d lmZej,eZGd d e dZGdde ZdgZy)zq Processor class for InstructBLIP. Largely copy of Blip2Processor with addition of a tokenizer for the Q-Former. N)ListUnion) BatchFeature) ImageInput)ProcessingKwargsProcessorMixinUnpack) AddedToken BatchEncodingPreTokenizedInput TextInput)logging) AutoTokenizerc *eZdZdddddddddd idZy)InstructBlipProcessorKwargsTFr) add_special_tokenspaddingstridereturn_overflowing_tokensreturn_special_tokens_maskreturn_offsets_mappingreturn_token_type_ids return_lengthverbose) text_kwargs images_kwargsN)__name__ __module__ __qualname__ _defaultsv/var/www/skyplay_api_hub/venv/lib/python3.12/site-packages/transformers/models/instructblip/processing_instructblip.pyrr&s0#').*/&+%*"   Ir$rF)totalc eZdZdZgdZdgZdZdZdZdfd Z dde de e e ee ee fd eed efd Zd Zd ZedZfdZefdZxZS)InstructBlipProcessora Constructs an InstructBLIP processor which wraps a BLIP image processor and a LLaMa/T5 tokenizer into a single processor. [`InstructBlipProcessor`] offers all the functionalities of [`BlipImageProcessor`] and [`AutoTokenizer`]. See the docstring of [`~BlipProcessor.__call__`] and [`~BlipProcessor.decode`] for more information. Args: image_processor (`BlipImageProcessor`): An instance of [`BlipImageProcessor`]. The image processor is a required input. tokenizer (`AutoTokenizer`): An instance of ['PreTrainedTokenizer`]. The tokenizer is a required input. qformer_tokenizer (`AutoTokenizer`): An instance of ['PreTrainedTokenizer`]. The Q-Former tokenizer is a required input. num_query_tokens (`int`, *optional*):" Number of tokens used by the Qformer as queries, should be same as in model's config. )image_processor tokenizerqformer_tokenizernum_query_tokens)BlipImageProcessorBlipImageProcessorFastrc t|ds2tddd|_|j|jgdn|j|_||_t ||||y)N image_tokenzFT) normalizedspecial)special_tokens)hasattrr r0 add_tokensr,super__init__)selfr)r*r+r,kwargs __class__s r%r7zInstructBlipProcessor.__init__Psby-0))tTD   $"2"2!3D I(44D  0 )5FGr$imagestextr9returnc | | td|jtfd|jji|}t }|ut |tr|g}n.t |tst |dts td|djdd}|j|fi|dddi} ||dd<|j{|yi} |jj|jz} |j| gt|zdd } | D]-} t| | | | Dcgc] \}}||z c}}| | </n| } |tj!d t#| | } |j%| |j&|fi|d}|jd |d <|jd|d<|'|j(|fi|d}|j%||Scc}}w)a This method uses [`BlipImageProcessor.__call__`] method to prepare image(s) for the model, and [`BertTokenizerFast.__call__`] to prepare text for the model. Please refer to the docstring of the above two methods for more information. Args: images (`ImageInput`): The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch tensor. Both channels-first and channels-last formats are supported. text (`TextInput`, `PreTokenizedInput`, `List[TextInput]`, `List[PreTokenizedInput]`): 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). Nz,You have to specify at least images or text.tokenizer_init_kwargsrzAInvalid input text. Please provide a string, or a list of stringsrreturn_tensorsF)rr@aAExpanding inputs for image tokens in InstructBLIP should be done in processing. Please follow instruction here (https://gist.github.com/zucchini-nlp/e9f20b054fa322f84ac9311d9ab67042) to update your InstructBLIP model. Using processors without these attributes in the config is deprecated and will throw an error in v4.50.) tensor_type input_idsqformer_input_idsattention_maskqformer_attention_maskr) ValueError _merge_kwargsrr* init_kwargsr isinstancestrlistpopr,r0contentlenziplogger warning_oncer updater+r))r8r;r<audiovideosr9 output_kwargsencodingr@_text_encoding text_encoding image_tokensimage_token_encodingk img_encoding txt_encodingqformer_text_encodingimage_encodings r%__call__zInstructBlipProcessor.__call__Ys?, >dlKL L*** ' "&.."<"<   >  $$vd+JtAw4L !dee+=9==>NPTUN+T^^DfM-4PfaefN=KM- ()9 :$$0V5G " #//77$:O:OO '+~~!NSY.5Y](6($(A;>>RST>UWefgWh:i(6L,%|3(M!$ !/ %''B*-^TM OOM *$:D$:$:4$`=Q^C_$` !,A,E,Ek,RH( )1F1J1JK[1\H- .  1T11&[M/>. )~.>>abc c NT2!#n>Q!R ../EF.@  OO " "#6 7').CFC  OO 34 4Or$c t||fi|}t|tr|d}t j|d}||_|S)Nrr+) subfolder)r6from_pretrainedrItuplerr+)clspretrained_model_name_or_pathr9 processorr+r:s r%r~z%InstructBlipProcessor.from_pretrainedsPG+,ITVT  i '!! I)99:Wcvw&7 #r$rh)NNNN)rr r!__doc__rv valid_kwargsimage_processor_classtokenizer_classqformer_tokenizer_classr7rrrr rr rrr`rbrfpropertyriru classmethodr~ __classcell__)r:s@r%r(r(7s$GJ&'LL%O-H"^b MMI0$y/4HYCZZ[M 45 M M`<6XX &r$r()rrptypingrrimage_processing_utilsr image_utilsrprocessing_utilsrr r tokenization_utils_baser r r rutilsrautor get_loggerrrPrr(__all__r#r$r%rsl 2%HH     H %"2%"eNeP # #r$