o ۷ij@s>ddlmZmZddlZddlmZmZddlmZm Z m Z m Z m Z m Z mZddlmZddlmZddlmZdd lmZdd lmZd d lmZmZeeZGd ddeZGdddeZ GdddeZ!Gddde Z"Gddde Z#Gddde Z$Gddde Z%Gddde Z&GdddeZ'gdZ(dS) )OptionalUnionN)InstructBlipQFormerConfigInstructBlipVisionConfig)$InstructBlipForConditionalGeneration/InstructBlipForConditionalGenerationModelOutputInstructBlipModelInstructBlipPreTrainedModelInstructBlipQFormerModelInstructBlipVisionModelTransformersKwargs)PretrainedConfig)FlashAttentionKwargs)!MODEL_FOR_CAUSAL_LM_MAPPING_NAMES)Unpack)logging)CONFIG_MAPPING AutoConfigc@ eZdZdS)InstructBlipVideoVisionConfigN__name__ __module__ __qualname__rru/home/ubuntu/vllm_env/lib/python3.10/site-packages/transformers/models/instructblipvideo/modular_instructblipvideo.pyr-rc@r)InstructBlipVideoQFormerConfigNrrrrrr1rrcs^eZdZdZdZddiZeeedZ     dfdd Z e d ed ed e fd dZ ZS)InstructBlipVideoConfiga [`InstructBlipVideoConfig`] is the configuration class to store the configuration of a [`InstructBlipVideoForConditionalGeneration`]. It is used to instantiate a Instructblipvideo model according to the specified arguments, defining the vision model, Q-Former model and language model configs. Instantiating a configuration with the defaults will yield a similar configuration to that of the Instructblipvideo [Salesforce/instruct-blip-flan-t5](https://huggingface.co/Salesforce/instruct-blip-flan-t5) architecture. Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the documentation from [`PretrainedConfig`] for more information. Args: vision_config (`dict`, *optional*): Dictionary of configuration options used to initialize [`InstructBlipVideoVisionConfig`]. qformer_config (`dict`, *optional*): Dictionary of configuration options used to initialize [`InstructBlipVideoQFormerConfig`]. text_config (`dict`, *optional*): Dictionary of configuration options used to initialize any [`PretrainedConfig`]. num_query_tokens (`int`, *optional*, defaults to 32): The number of query tokens passed through the Transformer. video_token_index (`int`, *optional*): Token index of special video token. kwargs (*optional*): Dictionary of keyword arguments. Example: ```python >>> from transformers import ( ... InstructBlipVideoVisionConfig, ... InstructBlipVideoQFormerConfig, ... OPTConfig, ... InstructBlipVideoConfig, ... InstructBlipVideoForConditionalGeneration, ... ) >>> # Initializing a InstructBlipVideoConfig with Salesforce/instruct-blip-flan-t5 style configuration >>> configuration = InstructBlipVideoConfig() >>> # Initializing a InstructBlipVideoForConditionalGeneration (with random weights) from the Salesforce/instruct-blip-flan-t5 style configuration >>> model = InstructBlipVideoForConditionalGeneration(configuration) >>> # Accessing the model configuration >>> configuration = model.config >>> # We can also initialize a InstructBlipVideoConfig from a InstructBlipVideoVisionConfig, InstructBlipVideoQFormerConfig and any PretrainedConfig >>> # Initializing Instructblipvideo vision, Instructblipvideo Q-Former and language model configurations >>> vision_config = InstructBlipVideoVisionConfig() >>> qformer_config = InstructBlipVideoQFormerConfig() >>> text_config = OPTConfig() >>> config = InstructBlipVideoConfig.from_text_vision_configs(vision_config, qformer_config, text_config) ```instructblipvideovideo_token_idvideo_token_index) text_configqformer_config vision_configN c stjdi||duri}td|duri}td|dur*i}tdtdi||_tdi||_|dd}t |di||_ ||_ ||_ |jj |j_|j jtv|_d|_d|_dS) NzZvision_config is None. initializing the InstructBlipVideoVisionConfig with default values.z\qformer_config is None. Initializing the InstructBlipVideoQFormerConfig with default values.zTtext_config is None. Initializing the text config with default values (`OPTConfig`). model_typeoptg?g{Gz?r)super__init__loggerinforr&rr%getrr$num_query_tokensr# hidden_sizeencoder_hidden_sizer(ruse_decoder_only_language_modelinitializer_factorinitializer_range)selfr&r%r$r/r#kwargstext_model_type __class__rrr+ws(       z InstructBlipVideoConfig.__init__r&r%r$cKs"|d|||d|S)a  Instantiate a [`InstructBlipVideoConfig`] (or a derived class) from a InstructBlipVideo vision model, Q-Former and language model configurations. Returns: [`InstructBlipVideoConfig`]: An instance of a configuration object )r&r%r$Nr)to_dict)clsr&r%r$r6rrr from_vision_qformer_text_configssz8InstructBlipVideoConfig.from_vision_qformer_text_configs)NNNr'N)rrr__doc__r( attribute_maprrr sub_configsr+ classmethodrr< __classcell__rrr8rr 5s.7#r c@r) InstructBlipVideoPreTrainedModelNrrrrrrBrrBc@r)InstructBlipVideoVisionModelNrrrrrrCrrCc@r)InstructBlipVideoQFormerModelNrrrrrrDrrDc@r)4InstructBlipVideoForConditionalGenerationModelOutputNrrrrrrErrEc!@seZdZ           ddejdejdeejdeejdeejdeejd eejd eejd eed eed eededeede e de e e ffddZdS)InstructBlipVideoModelNF pixel_valuesqformer_input_idsqformer_attention_mask input_idsattention_maskdecoder_input_idsdecoder_attention_mask inputs_embedsoutput_attentionsoutput_hidden_states return_dictinterpolate_pos_encoding use_cacher6returnc Ks:| dur| n|jj} |j\}}}}}||||||}|j|| | | | d}|d}tj|ddtj|j d}|j |jddd}tj|ddtj|j d}|dur^t |}|j |dd}|j |dd}tj||gdd}|j|||||| | | d}|dddd|dddf}||}|||jj|d}|dur|j|}||jjk}|durt |}n||tj|jjtj|j dk}|d}|d||j }||j |j}|||}|jjr|jd ||| | | | d|}n|jd ||||| | | | d |}t|||d S) N)rGrOrPrQrRrdtypedevicedim)rJrK query_embedsencoder_hidden_statesencoder_attention_maskrOrPrQrNrKrOrPrQrS)rNrKrLrMrOrPrQrS)vision_outputsqformer_outputslanguage_model_outputsr)configuse_return_dictshapereshape vision_modeltorchonessizelongrX query_tokensexpand ones_likerepeat_interleavecatqformerlanguage_projectionr/language_modelget_input_embeddingsr"tensorall unsqueeze expand_astorWmasked_scatterr2rE)r5rGrHrIrJrKrLrMrNrOrPrQrRrSr6 batch_sizeframeschannelheightwidthr` image_embedsimage_attention_maskrlquery_attention_mask query_outputs query_outputlanguage_model_inputsspecial_image_maskoutputsrrrforwards   $         zInstructBlipVideoModel.forward) NNNNNNNNNFN)rrrrh FloatTensorr LongTensorTensorboolrrrtuplerErrrrrrFsV      rFc#@seZdZ   ddejdejdeejdeedeef dd Z   ddejdejdeejdeedeef d d Z d ejd ejfddZ            ddejdejdeejd eejdeejdeejdeejd eejdeedeedeejdeededeede e de eeff ddZe      ddejdeejdeejd eejdeejd eejdedejfddZdS) )InstructBlipVideoForConditionalGenerationNFrGrHrIrRrQcCs>|j\}}}} } ||||| | }|j||dd} | d} tj| ddtj| jd} |j | jddd}tj|ddtj| jd}|durRt |}|j |dd}|j |dd}tj ||gdd}|j |||| | dd }|dddd|dddf}||}|||jj|d}|r|| |fS|S) a$ Encodes images into continuous embeddings that can be forwarded to the language model. Args: pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, image_size, image_size)`): The tensors corresponding to the input images. T)rGrRrQrNrUrVrYr[)rJrKr\r]r^rQ)rerfrgrhrirjrkrXrlrmrnrorprqrrrcr/)r5rGrHrIrRrQr{r|r}r~rr`rrrlrrrrrrrget_video_features0s<   $  zInstructBlipVideoForConditionalGeneration.get_placeholder_maskrKrLrMrOrPlabelsrSr6rTc Ks\| dur| n|jj} |j|||| dd\}}}| s|n|}| s%|n|}|dur1||}|dur:t|}||j|j }|j ||d}| ||}|jj r|j d ||| | | |d|}| rg|jn|d}d}| dur|jd || |jjjd|}n$|j d ||||| | | | |d |}| r|jn|d}| r|jn|d }t|||||d S) a qformer_input_ids (`torch.LongTensor` of shape (batch_size, sequence_length)): The sequence used as a prompt to be fed to the Q-Former module. qformer_attention_mask (`torch.LongTensor` of shape (batch_size, sequence_length), *optional*): Mask to avoid performing attention on padding token indices. Examples: ```python >>> from transformers import InstructBlipVideoProcessor, InstructBlipVideoForConditionalGeneration >>> import torch >>> from huggingface_hub import hf_hub_download >>> import av >>> import numpy as np >>> def read_video_pyav(container, indices): ... ''' ... Decode the video with PyAV decoder. ... Args: ... container (`av.container.input.InputContainer`): PyAV container. ... indices (`list[int]`): List of frame indices to decode. ... Returns: ... result (np.ndarray): np array of decoded frames of shape (num_frames, height, width, 3). ... ''' ... frames = [] ... container.seek(0) ... start_index = indices[0] ... end_index = indices[-1] ... for i, frame in enumerate(container.decode(video=0)): ... if i > end_index: ... break ... if i >= start_index and i in indices: ... frames.append(frame) ... return np.stack([x.to_ndarray(format="rgb24") for x in frames]) >>> model = InstructBlipVideoForConditionalGeneration.from_pretrained("Salesforce/instructblip-vicuna-7b", device_map="auto") >>> processor = InstructBlipVideoProcessor.from_pretrained("Salesforce/instructblip-vicuna-7b") >>> file_path = hf_hub_download( ... repo_id="nielsr/video-demo", filename="eating_spaghetti.mp4", repo_type="dataset" ... ) >>> container = av.open(file_path) >>> # sample uniformly 4 frames from the videWhy is this video funny?o >>> total_frames = container.streams.video[0].frames >>> indices = np.arange(0, total_frames, total_frames / 4).astype(int) >>> clip = read_video_pyav(container, indices) >>> prompt = "What is happening in the video?" >>> inputs = processor(text=prompt, images=clip, return_tensors="pt").to(model.device) >>> outputs = model.generate( ... **inputs, ... do_sample=False, ... num_beams=5, ... max_length=256, ... repetition_penalty=1.5, ... length_penalty=1.0, ... ) >>> generated_text = processor.batch_decode(outputs, skip_special_tokens=True)[0].strip() >>> print(generated_text) "A person is eating a bowl of pasta, and they are using a fork to eat it. The person is sitting at a table, and the plate of pasta is on the table in front" ```NTrHrIrRrQrNr_r)logitsr vocab_size) rNrKrLrMrOrPrQrrSr[)lossrr`rarbr)rcrdrto_tuplertrhrnryrXrWrrzr2rsr loss_functionr$rrrE)r5rGrHrIrJrKrLrMrNrOrPrrQrRrSr6rr`rrrrrrrrrsvQ        z1InstructBlipVideoForConditionalGeneration.forwardcKs t|dr ||jd} |j||||dd\} } } |durM|durG|jjg|jjd} | |jjjg}t j |gt j |j d}| | d}||}|durVt |}| |j |j} |j||d }||| }||d }|jjjsy||d <|jjd i||}|S) a Overrides `generate` function to be able to use the model as a conditional generator. Args: pixel_values (`torch.FloatTensor` of shape (batch_size, num_channels, height, width) or (batch_size, num_frames, num_channels, height, width)): Input images or videos to be processed. qformer_input_ids (`torch.LongTensor` of shape (batch_size, sequence_length), *optional*): The sequence used as a prompt to be fed to the Q-Former module. qformer_attention_mask (`torch.LongTensor` of shape (batch_size, sequence_length), *optional*): Mask to avoid performing attention on padding token indices. input_ids (`torch.LongTensor` of shape (batch_size, sequence_length), *optional*): The sequence used as a prompt for the generation. attention_mask (`torch.LongTensor` of shape (batch_size, sequence_length), *optional*): Mask to avoid performing attention on padding token indices. inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`): Embedded representation of the inputs. Should be float, not int tokens. interpolate_pos_encoding (`bool`, *optional*, defaults to `False`): Whether to interpolate the positional encoding of the image embeddings. Returns: captions (list): A list of strings of length batch_size * num_captions. hf_device_maprTrNrVr[r)rNrKrJr)hasattr_preprocess_acceleratererrcr#r/r$ bos_token_idrhrurkrXrepeatrtrnryrWrrzrsis_encoder_decodergenerate)r5rGrHrIrJrKrNrRgenerate_kwargsr{rr`r video_tokens start_tokensrinputsrrrrrs6 "        z2InstructBlipVideoForConditionalGeneration.generate)NFF) NNNNNNNNNNFN)NNNNNF)rrrrhrrrrrrrrr rrrErno_gradrrrrrr/s @          r)r rrrCrBrDrFr))typingrrrh;transformers.models.instructblip.configuration_instructbliprr6transformers.models.instructblip.modeling_instructbliprrrr r r r configuration_utilsrmodeling_flash_attention_utilsrmodels.auto.modeling_autorprocessing_utilsrutilsrautorr get_loggerrr,rrr rBrCrDrErFr__all__rrrrs.$     }m /