UhaHddlmZmZmZmZddlZddlZddlmZddlm Z ddl m Z ddl m Z ddlmZmZmZd d lmZmZd d lmZd d lmZmZmZmZmZej:eZGd deZ GddeZ!GddeZ"GddeZ#GddeZ$GddeZ%GddeZ&GddeZ'gdZ(y))ListOptionalTupleUnionN)nn) DynamicCache)FlashAttentionKwargs)Unpack)auto_docstringcan_return_tuplelogging)Idefics3ConfigIdefics3VisionConfig)Idefics3ImageProcessor)Idefics3BaseModelOutputWithPast Idefics3ForConditionalGeneration Idefics3ModelIdefics3PreTrainedModelIdefics3VisionTransformerceZdZdZdZy)SmolVLMVisionConfiga This is the configuration class to store the configuration of a [`SmolVLMVisionModel`]. It is used to instantiate a SmolVLM vision encoder according to the specified arguments, defining the model architecture. Instantiating a configuration with the defaults will yield a similar configuration to that of the SigLIP checkpoint [google/siglip-so400m-patch14-384](https://huggingface.co/google/siglip-so400m-patch14-384) used in SmolVLM [HuggingFaceTB/SmolVLM2-2.2B-Instruct](https://huggingface.co/HuggingFaceTB/SmolVLM2-2.2B-Instruct). Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the documentation from [`PretrainedConfig`] for more information. Args: hidden_size (`int`, *optional*, defaults to 1152): Dimensionality of the encoder layers and the pooler layer. intermediate_size (`int`, *optional*, defaults to 3072): Dimensionality of the "intermediate" (i.e., feed-forward) layer in the Transformer encoder. num_hidden_layers (`int`, *optional*, defaults to 12): Number of hidden layers in the Transformer encoder. num_attention_heads (`int`, *optional*, defaults to 16): Number of attention heads for each attention layer in the Transformer encoder. num_channels (`int`, *optional*, defaults to 3): Number of channels in the input images. image_size (`int`, *optional*, defaults to 224): The size (resolution) of each image. patch_size (`int`, *optional*, defaults to 32): The size (resolution) of each patch. hidden_act (`str` or `function`, *optional*, defaults to `"gelu_pytorch_tanh"`): The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`, `"relu"`, `"selu"` and `"gelu_new"` `"quick_gelu"` are supported. layer_norm_eps (`float`, *optional*, defaults to 1e-06): The epsilon used by the layer normalization layers. attention_dropout (`float`, *optional*, defaults to 0.0): The dropout ratio for the attention probabilities. initializer_range (`float`, *optional*, defaults to 0.02): The standard deviation of the truncated_normal_initializer for initializing all weight matrices. Example: ```python >>> from transformers.models.smolvlm.modeling_smolvlm import SmolVLMVisionTransformer >>> from transformers.models.smolvlm.configuration_smolvlm import SmolVLMVisionConfig >>> # Initializing a SmolVLMVisionConfig with google/siglip-so400m-patch14-384 style configuration >>> configuration = SmolVLMVisionConfig() >>> # Initializing a SmolVLMVisionTransformer (with random weights) from the google/siglip-so400m-patch14-384 style configuration >>> model = SmolVLMVisionTransformer(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```smolvlm_visionN__name__ __module__ __qualname____doc__ model_type}/var/www/catia.catastroantioquia-mas.com/valormas/lib/python3.12/site-packages/transformers/models/smolvlm/modular_smolvlm.pyrr(s1 f"Jr"rceZdZdZy)SmolVLMPreTrainedModelcDt|jd|jjj}t |t j t jfrY|jjjd||j%|jjjyyt |t jrf|jjjd||j2|jj|jjyyt |t jrJ|jjj!d|jjjyy)Ninitializer_range)meanstdg?)getattrconfigget_text_configr' isinstancerLinearConv2dweightdatanormal_biaszero_ Embedding padding_idx LayerNormfill_)selfmoduler*s r# _init_weightsz$SmolVLMPreTrainedModel._init_weightsasdkk#6 8S8S8U8g8gh fryy"))4 5 MM   & &CS & 9{{&   &&('  - MM   & &CS & 9!!- ""6#5#56<<>.  - MM   $ $S ) KK   " " $.r"N)rrrr<r!r"r#r%r%`s %r"r%c eZdZy)SmolVLMVisionTransformerNrrrr!r"r#r>r>qr"r>ceZdZdZdZy) SmolVLMConfiga This is the configuration class to store the configuration of a [`SmolVLMModel`]. It is used to instantiate a SmolVLM model according to the specified arguments, defining the model architecture. Instantiating a configuration with the defaults will yield a similar configuration to that of the model of the SmolVLM [HuggingFaceTB/SmolVLM2-2.2B-Instruct](https://huggingface.co/HuggingFaceTB/SmolVLM2-2.2B-Instruct) architecture. Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the documentation from [`PretrainedConfig`] for more information. Args: use_cache (`bool`, *optional*, defaults to `True`): Whether or not the model should cache the key/value pairs of the attention mechanism. Only relevant if `config.is_decoder=True`. image_token_id (`int`, *optional*, defaults to 128257): The id of the "image" token. tie_word_embeddings (`bool`, *optional*, defaults to `False`): Whether or not to tie the word embeddings with the token embeddings. vision_config (`IdeficsVisionConfig` or `dict`, *optional*, defaults to `IdeficsVisionConfig`): Custom vision config or dict for the vision tower text_config (`PretrainedConfig` or `dict`, *optional*, defaults to `LlamaConfig`): Custom text config or dict for the text model scale_factor (`int`, *optional*, defaults to 2): The scale factor for the image encoder. pad_token_id (`int`, *optional*, defaults to 128002): The id of the padding token. Example: ```python >>> from transformers import SmolVLMModel, SmolVLMConfig >>> # Initializing configuration >>> configuration = SmolVLMConfig() >>> # Initializing a model from the configuration >>> model = SmolVLMModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```smolvlmNrr!r"r#rBrBus# JJr"rBc eZdZy)SmolVLMImageProcessorNr?r!r"r#rErEr@r"rEc eZdZy)SmolVLMBaseModelOutputWithPastNr?r!r"r#rGrGr@r"rGc#4eZdZdZdej dej dej fdZddejdej fd Z e e d  dde ej d e ej d e ej de e ejde ejde ejde ejde ejde ede ede ede ede ej deedeeeffdZy) SmolVLMModelz A subclass of Idefics3Model. We do *not* remove or block the call to inputs_merger in forward. Instead, we override inputs_merger here with custom logic. input_ids inputs_embedsimage_hidden_statescH|j\}}}||jk(}|jd}tj||zdk(s t d||z}tj jj|jddd} | dd} |jd} | dz |z} | dz |z} | jd| z}tj|}|||| |ddf||<tj|jd||}|S)NdimrzCAt least one sample has tokens not divisible by patch_size.)rNr)value) shapeimage_token_idsumtorchall ValueErrorr functionalpadcumsum unsqueeze zeros_likewhere)r:rJrKrL_ patch_size image_masknum_image_tokensblocks_per_sampleoffsets block_offsetrow_cum chunk_idx local_idx block_idx image_embeds merged_embedss r# inputs_mergerzSmolVLMModel.inputs_mergers:/44:q$"5"55 %>>a>0yy)J6!;<bc c, :((%%))*;*B*Bq*B*I6YZ)[s| ###+q[Z/ q[J.  **1- 9 '' 6 #6y7LiXbNcef7f#g Z  J$8$8$P>P; J fe(|((j)@Z?l((+?jj#**$ $=#7#<# (batch_size, max_num_images, 3, max_heights, max_widths) where max_num_images is the maximum number of images among the batch_size samples in the batch. Padding images are not needed beyond padding the pixel_values at the entrance of the model. For efficiency, we only pass through the vision_model's forward the real images by discarding the padding images i.e. pixel_values of size (image_batch_size, 3, height, width) where image_batch_size would be 7 when num_images_per_sample=[1, 3, 1, 2] and max_num_images would be 3. ) custom_introattention_mask position_idspast_key_values use_cacheoutput_attentionsoutput_hidden_states return_dictcache_positionkwargsreturnc | | n|jj} | | n|jj} | | n|jj} | | n|jj} |j r/|j jr| rtjdd} ||j\}}n||j\}}}n tdd}| r| t}|j}|||dk(r td|9|j j|j|j }| | td||j#||}n)|'|j|j$|j }|||j'|||}|j d ||||| | | d | d |}t)|j*|j,|j.|j0| S) NzZ`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...Fz5You have to specify either input_ids or inputs_embedsrzWWhen first calling the model, if input_embeds are passed, input_ids should not be None.zMYou cannot specify both pixel_values and image_hidden_states at the same time)rsrt)rJrKrLT) rKrrrrrrrr)rr hidden_states attentionsrLr!)r,rrruse_return_dicttraining text_modelgradient_checkpointinglogger warning_oncerSrXr get_seq_lengthget_input_embeddingstortrrsrlrGrrrr)r:rJrrrrKrmrnrLrrrrrrr seq_lengthr_past_seen_tokensoutputss r#forwardzSmolVLMModel.forwards>:2C1N-TXT_T_TqTq$8$D $++JjJj "+!6IDKKV $D>V 'tnV d^V !!1!12V -.V u44 5!V  V r"rIc(eZdZfdZfdZxZS)SmolVLMForConditionalGenerationct||t||_t j |j j|j jd|_ |jy)NF)r4) super__init__rImodelrr/ text_config hidden_size vocab_sizelm_head post_init)r:r, __class__s r#rz(SmolVLMForConditionalGeneration.__init__ZsS  !&) yy!3!3!?!?ASASA^A^ejk  r"c $t|di|y)a Example: ```python >>> import requests >>> import torch >>> from PIL import Image >>> from io import BytesIO >>> from transformers import AutoProcessor, AutoModelForImageTextToText >>> from transformers.image_utils import load_image >>> # Note that passing the image urls (instead of the actual pil images) to the processor is also possible >>> image1 = load_image("https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg") >>> image2 = load_image("https://cdn.britannica.com/59/94459-050-DBA42467/Skyline-Chicago.jpg") >>> image3 = load_image("https://cdn.britannica.com/68/170868-050-8DDE8263/Golden-Gate-Bridge-San-Francisco.jpg") >>> processor = AutoProcessor.from_pretrained("HuggingFaceTB/SmolVLM2-2.2B-Instruct") >>> model = AutoModelForImageTextToText.from_pretrained("HuggingFaceTB/SmolVLM2-2.2B-Instruct", torch_dtype=torch.bfloat16, device_map="auto") >>> # Create inputs >>> messages = [ ... { ... "role": "user", ... "content": [ ... {"type": "video", "path": path/to/video}, ... {"type": "text", "text": "What is happening in this video?"}, ... ] ... } ... ] >>> inputs = processor.apply_chat_template([messages], add_generation_prompt=True) >>> # Generate >>> generated_ids = model.generate(**inputs, max_new_tokens=256) >>> generated_texts = processor.batch_decode(generated_ids, skip_special_tokens=True) >>> print(generated_texts) ```Nr!)rr)r: super_kwargsrs r#rz'SmolVLMForConditionalGeneration.forward`sP ','r")rrrrr __classcell__)rs@r#rrYs ((((r"r)rrBrErr%rIr>))typingrrrrrVtorch.utils.checkpointr cache_utilsr modeling_flash_attention_utilsr processing_utilsr utilsr r ridefics3.configuration_idefics3rr"idefics3.image_processing_idefics3ridefics3.modeling_idefics3rrrrr get_loggerrrrr%r>rBrErGrIr__all__r!r"r#rs 0/ 'B&>>RG   H %5 .5 p%4%" 8 ' N' T 2  %D o =o d/(&F/(d r"