# Idefics3 ## Overview The Idefics3 model was proposed in [Building and better understanding vision-language models: insights and future directions](https://huggingface.co/papers/2408.12637) by Hugo Laurençon, Andrés Marafioti, Victor Sanh, and Léo Tronchon. Idefics3 is an adaptation of the Idefics2 model with three main differences: - It uses Llama3 for the text model. - It uses an updated processing logic for the images. - It removes the perceiver. The abstract from the paper is the following: *The field of vision-language models (VLMs), which take images and texts as inputs and output texts, is rapidly evolving and has yet to reach consensus on several key aspects of the development pipeline, including data, architecture, and training methods. This paper can be seen as a tutorial for building a VLM. We begin by providing a comprehensive overview of the current state-of-the-art approaches, highlighting the strengths and weaknesses of each, addressing the major challenges in the field, and suggesting promising research directions for underexplored areas. We then walk through the practical steps to build Idefics3-8B, a powerful VLM that significantly outperforms its predecessor Idefics2-8B, while being trained efficiently, exclusively on open datasets, and using a straightforward pipeline. These steps include the creation of Docmatix, a dataset for improving document understanding capabilities, which is 240 times larger than previously available datasets. We release the model along with the datasets created for its training.* ## Usage tips Input images are processed either by upsampling (if resizing is enabled) or at their original resolution. The resizing behavior depends on two parameters: do_resize and size. If `do_resize` is set to `True`, the model resizes images so that the longest edge is 4*364 pixels by default. The default resizing behavior can be customized by passing a dictionary to the `size` parameter. For example, `{"longest_edge": 4 * 364}` is the default, but you can change it to a different value if needed. Here’s how to control resizing and set a custom size: ```python image_processor = Idefics3ImageProcessor(do_resize=True, size={"longest_edge": 2 * 364}, max_image_size=364) ``` Additionally, the `max_image_size` parameter, which controls the size of each square patch the image is decomposed into, is set to 364 by default but can be adjusted as needed. After resizing (if applicable), the image processor decomposes the images into square patches based on the `max_image_size` parameter. This model was contributed by [amyeroberts](https://huggingface.co/amyeroberts) and [andimarafioti](https://huggingface.co/andito). ## Idefics3Config[[transformers.Idefics3Config]] #### transformers.Idefics3Config[[transformers.Idefics3Config]] [Source](https://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/configuration_idefics3.py#L108) This is the configuration class to store the configuration of a [Idefics3Model](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3Model). It is used to instantiate a Idefics3 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 Idefics3 [HuggingFaceM4/Idefics3-8B-Llama3](https://huggingface.co/HuggingFaceM4/Idefics3-8B-Llama3) architecture. Configuration objects inherit from [PreTrainedConfig](/docs/transformers/v5.2.0/en/main_classes/configuration#transformers.PreTrainedConfig) and can be used to control the model outputs. Read the documentation from [PreTrainedConfig](/docs/transformers/v5.2.0/en/main_classes/configuration#transformers.PreTrainedConfig) for more information. Example: ```python >>> from transformers import Idefics3Model, Idefics3Config >>> # Initializing configuration >>> configuration = Idefics3Config() >>> # Initializing a model from the configuration >>> model = Idefics3Model(configuration) >>> # Accessing the model configuration >>> configuration = model.config ``` **Parameters:** 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. ## Idefics3VisionConfig[[transformers.Idefics3VisionConfig]] #### transformers.Idefics3VisionConfig[[transformers.Idefics3VisionConfig]] [Source](https://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/configuration_idefics3.py#L23) This is the configuration class to store the configuration of a `Idefics3VisionModel`. It is used to instantiate a Idefics3 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-base-patch16-224](https://huggingface.co/google/siglip-base-patch16-224) used in the Idefics3 model [HuggingFaceM4/Idefics3-8B-Llama3](https://huggingface.co/HuggingFaceM4/Idefics3-8B-Llama3). Configuration objects inherit from [PreTrainedConfig](/docs/transformers/v5.2.0/en/main_classes/configuration#transformers.PreTrainedConfig) and can be used to control the model outputs. Read the documentation from [PreTrainedConfig](/docs/transformers/v5.2.0/en/main_classes/configuration#transformers.PreTrainedConfig) for more information. Example: ```python >>> from transformers.models.idefics3.modeling_idefics3 import Idefics3VisionTransformer >>> from transformers.models.idefics3.configuration_idefics3 import Idefics3VisionConfig >>> # Initializing a Idefics3VisionConfig with google/siglip-base-patch16-224 style configuration >>> configuration = Idefics3VisionConfig() >>> # Initializing a Idefics3VisionTransformer (with random weights) from the google/siglip-base-patch16-224 style configuration >>> model = Idefics3VisionTransformer(configuration) >>> # Accessing the model configuration >>> configuration = model.config ``` **Parameters:** 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. ## Idefics3VisionTransformer[[transformers.Idefics3VisionTransformer]] #### transformers.Idefics3VisionTransformer[[transformers.Idefics3VisionTransformer]] [Source](https://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/modeling_idefics3.py#L444) The Idefics3 Vision Transformer Model outputting raw image embedding. This model inherits from [PreTrainedModel](/docs/transformers/v5.2.0/en/main_classes/model#transformers.PreTrainedModel). Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.) This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior. **Parameters:** config ([Idefics3VisionConfig](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3VisionConfig)) : Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [from_pretrained()](/docs/transformers/v5.2.0/en/main_classes/model#transformers.PreTrainedModel.from_pretrained) method to load the model weights. ## Idefics3Model[[transformers.Idefics3Model]] #### transformers.Idefics3Model[[transformers.Idefics3Model]] [Source](https://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/modeling_idefics3.py#L522) Idefics3 model consisting of a SIGLIP vision encoder and Llama3 language decoder This model inherits from [PreTrainedModel](/docs/transformers/v5.2.0/en/main_classes/model#transformers.PreTrainedModel). Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.) This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior. forwardtransformers.Idefics3Model.forwardhttps://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/modeling_idefics3.py#L627[{"name": "input_ids", "val": ": torch.LongTensor | None = None"}, {"name": "attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "position_ids", "val": ": torch.LongTensor | None = None"}, {"name": "past_key_values", "val": ": transformers.cache_utils.Cache | None = None"}, {"name": "inputs_embeds", "val": ": torch.FloatTensor | None = None"}, {"name": "pixel_values", "val": ": torch.FloatTensor | None = None"}, {"name": "pixel_attention_mask", "val": ": torch.BoolTensor | None = None"}, {"name": "image_hidden_states", "val": ": torch.FloatTensor | None = None"}, {"name": "use_cache", "val": ": bool | None = None"}, {"name": "output_attentions", "val": ": bool | None = None"}, {"name": "output_hidden_states", "val": ": bool | None = None"}, {"name": "cache_position", "val": ": torch.LongTensor | None = None"}, {"name": "return_dict", "val": ": bool | None = None"}, {"name": "**kwargs", "val": ": typing_extensions.Unpack[transformers.modeling_flash_attention_utils.FlashAttentionKwargs]"}]- **input_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Indices of input sequence tokens in the vocabulary. Padding will be ignored by default. Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.2.0/en/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.2.0/en/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and [PreTrainedTokenizer.__call__()](/docs/transformers/v5.2.0/en/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__call__) for details. [What are input IDs?](../glossary#input-ids) - **attention_mask** (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*) -- Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`: - 1 for tokens that are **not masked**, - 0 for tokens that are **masked**. [What are attention masks?](../glossary#attention-mask) - **position_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, config.n_positions - 1]`. [What are position IDs?](../glossary#position-ids) - **past_key_values** (`~cache_utils.Cache`, *optional*) -- Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values` returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`. Only [Cache](/docs/transformers/v5.2.0/en/internal/generation_utils#transformers.Cache) instance is allowed as input, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). If no `past_key_values` are passed, [DynamicCache](/docs/transformers/v5.2.0/en/internal/generation_utils#transformers.DynamicCache) will be initialized by default. The model will output the same cache format that is fed as input. If `past_key_values` are used, the user is expected to input only unprocessed `input_ids` (those that don't have their past key value states given to this model) of shape `(batch_size, unprocessed_length)` instead of all `input_ids` of shape `(batch_size, sequence_length)`. - **inputs_embeds** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*) -- Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert `input_ids` indices into associated vectors than the model's internal embedding lookup matrix. - **pixel_values** (`torch.FloatTensor` of shape `(batch_size, num_channels, image_size, image_size)`, *optional*) -- The tensors corresponding to the input images. Pixel values can be obtained using [Idefics3ImageProcessorFast](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3ImageProcessorFast). See [Idefics3ImageProcessorFast.__call__()](/docs/transformers/v5.2.0/en/model_doc/fuyu#transformers.FuyuImageProcessor.__call__) for details ([Idefics3Processor](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3Processor) uses [Idefics3ImageProcessorFast](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3ImageProcessorFast) for processing images). - **pixel_attention_mask** (`torch.Tensor` of shape `(batch_size, image_size, image_size)`, *optional*) -- Mask to avoid performing attention on padding pixel indices. - **image_hidden_states** (`torch.FloatTensor` of shape `(batch_size, num_channels, image_size, image_size)`) -- The hidden states of the image encoder after modality projection. - **use_cache** (`bool`, *optional*) -- If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see `past_key_values`). - **output_attentions** (`bool`, *optional*) -- Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned tensors for more detail. - **output_hidden_states** (`bool`, *optional*) -- Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for more detail. - **cache_position** (`torch.LongTensor` of shape `(sequence_length)`, *optional*) -- Indices depicting the position of the input sequence tokens in the sequence. Contrarily to `position_ids`, this tensor is not affected by padding. It is used to update the cache in the correct position and to infer the complete sequence length. - **return_dict** (`bool`, *optional*) -- Whether or not to return a [ModelOutput](/docs/transformers/v5.2.0/en/main_classes/output#transformers.utils.ModelOutput) instead of a plain tuple.0`transformers.models.idefics3.modeling_idefics3.Idefics3BaseModelOutputWithPast` or `tuple(torch.FloatTensor)`A `transformers.models.idefics3.modeling_idefics3.Idefics3BaseModelOutputWithPast` or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([Idefics3Config](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3Config)) and inputs. - **last_hidden_state** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`) -- Sequence of hidden-states at the output of the last layer of the model. If `past_key_values` is used only the last hidden-state of the sequences of shape `(batch_size, 1, hidden_size)` is output. - **past_key_values** (`Cache`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`) -- It is a [Cache](/docs/transformers/v5.2.0/en/internal/generation_utils#transformers.Cache) instance. For more details, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). Contains pre-computed hidden-states (key and values in the self-attention blocks and optionally if `config.is_encoder_decoder=True` in the cross-attention blocks) that can be used (see `past_key_values` input) to speed up sequential decoding. - **hidden_states** (`tuple[torch.FloatTensor] | None.hidden_states`, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the model at the output of each layer plus the optional initial embedding outputs. - **attentions** (`tuple[torch.FloatTensor] | None.attentions`, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. - **image_hidden_states** (`tuple(torch.FloatTensor)`, *optional*) -- Tuple of `torch.FloatTensor` (one for the output of the image embeddings, `(batch_size, num_images, sequence_length, hidden_size)`. image_hidden_states of the model produced by the vision encoder Inputs fed to the model can have an arbitrary number of images. To account for this, pixel_values fed to the model have image padding -> (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. **Parameters:** config ([Idefics3Config](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3Config)) : Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [from_pretrained()](/docs/transformers/v5.2.0/en/main_classes/model#transformers.PreTrainedModel.from_pretrained) method to load the model weights. **Returns:** ``transformers.models.idefics3.modeling_idefics3.Idefics3BaseModelOutputWithPast` or `tuple(torch.FloatTensor)`` A `transformers.models.idefics3.modeling_idefics3.Idefics3BaseModelOutputWithPast` or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([Idefics3Config](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3Config)) and inputs. - **last_hidden_state** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`) -- Sequence of hidden-states at the output of the last layer of the model. If `past_key_values` is used only the last hidden-state of the sequences of shape `(batch_size, 1, hidden_size)` is output. - **past_key_values** (`Cache`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`) -- It is a [Cache](/docs/transformers/v5.2.0/en/internal/generation_utils#transformers.Cache) instance. For more details, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). Contains pre-computed hidden-states (key and values in the self-attention blocks and optionally if `config.is_encoder_decoder=True` in the cross-attention blocks) that can be used (see `past_key_values` input) to speed up sequential decoding. - **hidden_states** (`tuple[torch.FloatTensor] | None.hidden_states`, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the model at the output of each layer plus the optional initial embedding outputs. - **attentions** (`tuple[torch.FloatTensor] | None.attentions`, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. - **image_hidden_states** (`tuple(torch.FloatTensor)`, *optional*) -- Tuple of `torch.FloatTensor` (one for the output of the image embeddings, `(batch_size, num_images, sequence_length, hidden_size)`. image_hidden_states of the model produced by the vision encoder #### get_image_features[[transformers.Idefics3Model.get_image_features]] [Source](https://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/modeling_idefics3.py#L575) **Parameters:** pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, image_size, image_size)`) : The tensors corresponding to the input images. pixel_attention_mask (`torch.LongTensor`, *optional*) : The attention mask indicating padded regions in the image. **Returns:** `[transformers.modeling_outputs.BaseModelOutputWithPooling](/docs/transformers/v5.2.0/en/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPooling) or `tuple(torch.FloatTensor)`` A [transformers.modeling_outputs.BaseModelOutputWithPooling](/docs/transformers/v5.2.0/en/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPooling) or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([Idefics3Config](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3Config)) and inputs. - **last_hidden_state** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`) -- Sequence of hidden-states at the output of the last layer of the model. - **pooler_output** (`torch.FloatTensor` of shape `(batch_size, hidden_size)`) -- Last layer hidden-state of the first token of the sequence (classification token) after further processing through the layers used for the auxiliary pretraining task. E.g. for BERT-family of models, this returns the classification token after processing through a linear layer and a tanh activation function. The linear layer weights are trained from the next sentence prediction (classification) objective during pretraining. - **hidden_states** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the model at the output of each layer plus the optional initial embedding outputs. - **attentions** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. ## Idefics3ForConditionalGeneration[[transformers.Idefics3ForConditionalGeneration]] #### transformers.Idefics3ForConditionalGeneration[[transformers.Idefics3ForConditionalGeneration]] [Source](https://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/modeling_idefics3.py#L735) The Idefics3 Model with a language modeling head. It is made up a SigLIP vision encoder, with a language modeling head on top. This model inherits from [PreTrainedModel](/docs/transformers/v5.2.0/en/main_classes/model#transformers.PreTrainedModel). Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.) This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior. forwardtransformers.Idefics3ForConditionalGeneration.forwardhttps://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/modeling_idefics3.py#L775[{"name": "input_ids", "val": ": torch.LongTensor | None = None"}, {"name": "attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "position_ids", "val": ": torch.LongTensor | None = None"}, {"name": "past_key_values", "val": ": transformers.cache_utils.Cache | None = None"}, {"name": "inputs_embeds", "val": ": torch.FloatTensor | None = None"}, {"name": "pixel_values", "val": ": torch.FloatTensor | None = None"}, {"name": "pixel_attention_mask", "val": ": torch.BoolTensor | None = None"}, {"name": "image_hidden_states", "val": ": torch.FloatTensor | None = None"}, {"name": "labels", "val": ": torch.LongTensor | None = None"}, {"name": "use_cache", "val": ": bool | None = None"}, {"name": "output_attentions", "val": ": bool | None = None"}, {"name": "output_hidden_states", "val": ": bool | None = None"}, {"name": "cache_position", "val": ": torch.LongTensor | None = None"}, {"name": "return_dict", "val": ": bool | None = None"}, {"name": "logits_to_keep", "val": ": int | torch.Tensor = 0"}, {"name": "**kwargs", "val": ": typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs]"}]- **input_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Indices of input sequence tokens in the vocabulary. Padding will be ignored by default. Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.2.0/en/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.2.0/en/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and [PreTrainedTokenizer.__call__()](/docs/transformers/v5.2.0/en/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__call__) for details. [What are input IDs?](../glossary#input-ids) - **attention_mask** (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*) -- Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`: - 1 for tokens that are **not masked**, - 0 for tokens that are **masked**. [What are attention masks?](../glossary#attention-mask) - **position_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, config.n_positions - 1]`. [What are position IDs?](../glossary#position-ids) - **past_key_values** (`~cache_utils.Cache`, *optional*) -- Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values` returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`. Only [Cache](/docs/transformers/v5.2.0/en/internal/generation_utils#transformers.Cache) instance is allowed as input, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). If no `past_key_values` are passed, [DynamicCache](/docs/transformers/v5.2.0/en/internal/generation_utils#transformers.DynamicCache) will be initialized by default. The model will output the same cache format that is fed as input. If `past_key_values` are used, the user is expected to input only unprocessed `input_ids` (those that don't have their past key value states given to this model) of shape `(batch_size, unprocessed_length)` instead of all `input_ids` of shape `(batch_size, sequence_length)`. - **inputs_embeds** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*) -- Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert `input_ids` indices into associated vectors than the model's internal embedding lookup matrix. - **pixel_values** (`torch.FloatTensor` of shape `(batch_size, num_channels, image_size, image_size)`, *optional*) -- The tensors corresponding to the input images. Pixel values can be obtained using [Idefics3ImageProcessorFast](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3ImageProcessorFast). See [Idefics3ImageProcessorFast.__call__()](/docs/transformers/v5.2.0/en/model_doc/fuyu#transformers.FuyuImageProcessor.__call__) for details ([Idefics3Processor](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3Processor) uses [Idefics3ImageProcessorFast](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3ImageProcessorFast) for processing images). - **pixel_attention_mask** (`torch.Tensor` of shape `(batch_size, image_size, image_size)`, *optional*) -- Mask to avoid performing attention on padding pixel indices. - **image_hidden_states** (`torch.FloatTensor` of shape `(batch_size, num_channels, image_size, image_size)`) -- The hidden states of the image encoder after modality projection. - **labels** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) -- Labels for computing the masked language modeling loss. Indices should either be in `[0, ..., config.vocab_size]` or `model.image_token_id` (where `model` is your instance of `Idefics3ForConditionalGeneration`). Tokens with indices set to `model.image_token_id` are ignored (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`. - **use_cache** (`bool`, *optional*) -- If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see `past_key_values`). - **output_attentions** (`bool`, *optional*) -- Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned tensors for more detail. - **output_hidden_states** (`bool`, *optional*) -- Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for more detail. - **cache_position** (`torch.LongTensor` of shape `(sequence_length)`, *optional*) -- Indices depicting the position of the input sequence tokens in the sequence. Contrarily to `position_ids`, this tensor is not affected by padding. It is used to update the cache in the correct position and to infer the complete sequence length. - **return_dict** (`bool`, *optional*) -- Whether or not to return a [ModelOutput](/docs/transformers/v5.2.0/en/main_classes/output#transformers.utils.ModelOutput) instead of a plain tuple. - **logits_to_keep** (`Union[int, torch.Tensor]`, *optional*, defaults to `0`) -- If an `int`, compute logits for the last `logits_to_keep` tokens. If `0`, calculate logits for all `input_ids` (special case). Only last token logits are needed for generation, and calculating them only for that token can save memory, which becomes pretty significant for long sequences or large vocabulary size. If a `torch.Tensor`, must be 1D corresponding to the indices to keep in the sequence length dimension. This is useful when using packed tensor format (single dimension for batch and sequence length).0`transformers.models.idefics3.modeling_idefics3.Idefics3CausalLMOutputWithPast` or `tuple(torch.FloatTensor)`A `transformers.models.idefics3.modeling_idefics3.Idefics3CausalLMOutputWithPast` or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([Idefics3Config](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3Config)) and inputs. - **loss** (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided) -- Language modeling loss (for next-token prediction). - **logits** (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`) -- Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax). - **past_key_values** (`Cache`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`) -- It is a [Cache](/docs/transformers/v5.2.0/en/internal/generation_utils#transformers.Cache) instance. For more details, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see `past_key_values` input) to speed up sequential decoding. - **hidden_states** (`tuple[torch.FloatTensor] | None.hidden_states`, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the model at the output of each layer plus the optional initial embedding outputs. - **attentions** (`tuple[torch.FloatTensor] | None.attentions`, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. - **image_hidden_states** (`tuple(torch.FloatTensor)`, *optional*) -- Tuple of `torch.FloatTensor` (one for the output of the image embeddings, `(batch_size, num_images, sequence_length, hidden_size)`. image_hidden_states of the model produced by the vision encoder The [Idefics3ForConditionalGeneration](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3ForConditionalGeneration) forward method, overrides the `__call__` special method. Although the recipe for forward pass needs to be defined within this function, one should call the `Module` instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them. Example: ```python >>> 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("HuggingFaceM4/Idefics3-8B-Llama3") >>> model = AutoModelForImageTextToText.from_pretrained("HuggingFaceM4/Idefics3-8B-Llama3", dtype=torch.bfloat16, device_map="auto") >>> # Create inputs >>> messages = [ ... { ... "role": "user", ... "content": [ ... {"type": "image"}, ... {"type": "text", "text": "In this image, we can see the city of New York, and more specifically the Statue of Liberty."}, ... {"type": "image"}, ... {"type": "text", "text": "What can we see in this image?"}, ... ] ... }, ... { ... "role": "user", ... "content": [ ... {"type": "image"}, ... {"type": "text", "text": "In which city is that bridge located?"}, ... ] ... } ... ] >>> prompts = [processor.apply_chat_template([message], add_generation_prompt=True) for message in messages] >>> images = [[image1, image2], [image3]] >>> inputs = processor(text=prompts, images=images, padding=True, return_tensors="pt").to(model.device) >>> # Generate >>> generated_ids = model.generate(**inputs, max_new_tokens=256) >>> generated_texts = processor.batch_decode(generated_ids, skip_special_tokens=True) >>> print(generated_texts[0]) Assistant: There are buildings, trees, lights, and water visible in this image. >>> print(generated_texts[1]) Assistant: The bridge is in San Francisco. ``` **Parameters:** config ([Idefics3ForConditionalGeneration](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3ForConditionalGeneration)) : Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [from_pretrained()](/docs/transformers/v5.2.0/en/main_classes/model#transformers.PreTrainedModel.from_pretrained) method to load the model weights. **Returns:** ``transformers.models.idefics3.modeling_idefics3.Idefics3CausalLMOutputWithPast` or `tuple(torch.FloatTensor)`` A `transformers.models.idefics3.modeling_idefics3.Idefics3CausalLMOutputWithPast` or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([Idefics3Config](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3Config)) and inputs. - **loss** (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided) -- Language modeling loss (for next-token prediction). - **logits** (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`) -- Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax). - **past_key_values** (`Cache`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`) -- It is a [Cache](/docs/transformers/v5.2.0/en/internal/generation_utils#transformers.Cache) instance. For more details, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache). Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see `past_key_values` input) to speed up sequential decoding. - **hidden_states** (`tuple[torch.FloatTensor] | None.hidden_states`, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the model at the output of each layer plus the optional initial embedding outputs. - **attentions** (`tuple[torch.FloatTensor] | None.attentions`, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. - **image_hidden_states** (`tuple(torch.FloatTensor)`, *optional*) -- Tuple of `torch.FloatTensor` (one for the output of the image embeddings, `(batch_size, num_images, sequence_length, hidden_size)`. image_hidden_states of the model produced by the vision encoder #### get_image_features[[transformers.Idefics3ForConditionalGeneration.get_image_features]] [Source](https://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/modeling_idefics3.py#L758) Example: ```python >>> from PIL import Image >>> from transformers import AutoProcessor, Idefics3ForConditionalGeneration >>> model = Idefics3ForConditionalGeneration.from_pretrained("HuggingFaceM4/Idefics3-8B-Llama3") >>> processor = AutoProcessor.from_pretrained("HuggingFaceM4/Idefics3-8B-Llama3") >>> messages = [ ... { ... "role": "user", "content": [ ... {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/pipeline-cat-chonk.jpeg"}, ... {"type": "text", "text": "Where is the cat standing?"}, ... ] ... }, ... ] >>> inputs = processor.apply_chat_template( ... messages, ... tokenize=True, ... return_dict=True, ... return_tensors="pt", ... add_generation_prompt=True ... ) >>> # Generate >>> generate_ids = model.generate(**inputs) >>> processor.batch_decode(generate_ids, skip_special_tokens=True)[0] ``` **Parameters:** pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, image_size, image_size)`) : The tensors corresponding to the input images. pixel_attention_mask (`torch.LongTensor`, *optional*) : The attention mask indicating padded regions in the image. **Returns:** `[transformers.modeling_outputs.BaseModelOutputWithPooling](/docs/transformers/v5.2.0/en/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPooling) or `tuple(torch.FloatTensor)`` A [transformers.modeling_outputs.BaseModelOutputWithPooling](/docs/transformers/v5.2.0/en/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPooling) or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([Idefics3Config](/docs/transformers/v5.2.0/en/model_doc/idefics3#transformers.Idefics3Config)) and inputs. - **last_hidden_state** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`) -- Sequence of hidden-states at the output of the last layer of the model. - **pooler_output** (`torch.FloatTensor` of shape `(batch_size, hidden_size)`) -- Last layer hidden-state of the first token of the sequence (classification token) after further processing through the layers used for the auxiliary pretraining task. E.g. for BERT-family of models, this returns the classification token after processing through a linear layer and a tanh activation function. The linear layer weights are trained from the next sentence prediction (classification) objective during pretraining. - **hidden_states** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`. Hidden-states of the model at the output of each layer plus the optional initial embedding outputs. - **attentions** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length, sequence_length)`. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. ## Idefics3ImageProcessor[[transformers.Idefics3ImageProcessor]] #### transformers.Idefics3ImageProcessor[[transformers.Idefics3ImageProcessor]] [Source](https://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/image_processing_idefics3.py#L268) Constructs a Idefics3 image processor. preprocesstransformers.Idefics3ImageProcessor.preprocesshttps://github.com/huggingface/transformers/blob/v5.2.0/src/transformers/models/idefics3/image_processing_idefics3.py#L620[{"name": "images", "val": ": typing.Union[ForwardRef('PIL.Image.Image'), numpy.ndarray, ForwardRef('torch.Tensor'), list['PIL.Image.Image'], list[numpy.ndarray], list['torch.Tensor']]"}, {"name": "do_convert_rgb", "val": ": bool | None = None"}, {"name": "do_resize", "val": ": bool | None = None"}, {"name": "size", "val": ": dict[str, int] | None = None"}, {"name": "resample", "val": ": PIL.Image.Resampling | None = None"}, {"name": "do_image_splitting", "val": ": bool | None = None"}, {"name": "do_rescale", "val": ": bool | None = None"}, {"name": "max_image_size", "val": ": dict[str, int] | None = None"}, {"name": "rescale_factor", "val": ": float | None = None"}, {"name": "do_normalize", "val": ": bool | None = None"}, {"name": "image_mean", "val": ": float | list[float] | None = None"}, {"name": "image_std", "val": ": float | list[float] | None = None"}, {"name": "do_pad", "val": ": bool | None = None"}, {"name": "return_tensors", "val": ": str | transformers.utils.generic.TensorType | None = None"}, {"name": "return_row_col_info", "val": ": bool = False"}, {"name": "data_format", "val": ": transformers.image_utils.ChannelDimension | None = "}, {"name": "input_data_format", "val": ": str | transformers.image_utils.ChannelDimension | None = None"}]- **images** (`ImageInput`) -- A list of images to preprocess. - **do_convert_rgb** (`bool`, *optional*, defaults to `self.do_convert_rgb`) -- Whether to convert the image to RGB. - **do_resize** (`bool`, *optional*, defaults to `self.do_resize`) -- Whether to resize the image. - **size** (`dict[str, int]`, *optional*, defaults to `self.size`) -- Size of the image after resizing. With the longest edge resized to keep the input aspect ratio. - **resample** (`int`, *optional*, defaults to `self.resample`) -- Resampling filter to use if resizing the image. This can be one of the enum `PILImageResampling`. Only has an effect if `do_resize` is set to `True`. - **do_image_splitting** (`bool`, *optional*, defaults to `self.do_image_splitting`) -- Whether to split the image into sub-images concatenated with the original image. They are split into patches such that each patch has a size of `max_image_size["height"]` x `max_image_size["width"]`. - **max_image_size** (`Dict`, *optional*, defaults to `self.max_image_size`) -- Maximum resolution of the images. If the image is larger than this size, the image is split into patches. - **do_rescale** (`bool`, *optional*, defaults to `self.do_rescale`) -- Whether to rescale the image. - **rescale_factor** (`float`, *optional*, defaults to `self.rescale_factor`) -- Rescale factor to rescale the image by if `do_rescale` is set to `True`. - **do_normalize** (`bool`, *optional*, defaults to `self.do_normalize`) -- Whether to normalize the image. - **image_mean** (`float` or `list[float]`, *optional*, defaults to `self.image_mean`) -- Image mean to use for normalization. Only has an effect if `do_normalize` is set to `True`. - **image_std** (`float` or `list[float]`, *optional*, defaults to `self.image_std`) -- Image standard deviation to use for normalization. Only has an effect if `do_normalize` is set to `True`. - **do_pad** (`bool`, *optional*, defaults to `self.do_pad`) -- Whether or not to pad the images to the largest height and width in the batch. - **return_tensors** (`str` or `TensorType`, *optional*) -- The type of tensors to return. Can be one of: - Unset: Return a list of `np.ndarray`. - `TensorType.PYTORCH` or `'pt'`: Return a batch of type `torch.Tensor`. - `TensorType.NUMPY` or `'np'`: Return a batch of type `np.ndarray`. - **return_row_col_info** (`bool`, *optional*, default to `False`) -- Whether to return the number of rows and columns of the split images. This is used for the `Idefics3Processor` to generate prompt strings based on the number of rows and columns. - **data_format** (`ChannelDimension` or `str`, *optional*, defaults to `ChannelDimension.FIRST`) -- The channel dimension format for the output image. Can be one of: - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. - Unset: Use the channel dimension format of the input image. - **input_data_format** (`ChannelDimension` or `str`, *optional*) -- The channel dimension format for the input image. If unset, the channel dimension format is inferred from the input image. Can be one of: - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. - `"none"` or `ChannelDimension.NONE`: image in (height, width) format.0 Preprocess a batch of images. **Parameters:** do_convert_rgb (`bool`, *optional*, defaults to `True`) : Whether to convert the image to RGB. This is useful if the input image is of a different format e.g. RGBA. Only has an effect if the input image is in the PIL format. do_resize (`bool`, *optional*, defaults to `True`) : Whether to resize the image. The longest edge of the image is resized to be tokens per image in the input. This parameter is used to build the string from the input prompt and image tokens and should match the value the model used. It is computed as: image_seq_len = int(((image_size // patch_size) ** 2) / (scale_factor**2)) chat_template (`str`, *optional*) : A Jinja template to convert lists of messages in a chat into a tokenizable string. **Returns:** ```` - **data** (`dict`, *optional*) -- Dictionary of lists/arrays/tensors returned by the `__call__`/`encode_plus`/`batch_encode_plus` methods ('input_ids', 'attention_mask', etc.). - **encoding** (`tokenizers.Encoding` or `Sequence[tokenizers.Encoding]`, *optional*) -- If the tokenizer is a fast tokenizer which outputs additional information like mapping from word/character space to token space the `tokenizers.Encoding` instance or list of instance (for batches) hold this information. - **tensor_type** (`Union[None, str, TensorType]`, *optional*) -- You can give a tensor_type here to convert the lists of integers in PyTorch/Numpy Tensors at initialization. - **prepend_batch_axis** (`bool`, *optional*, defaults to `False`) -- Whether or not to add a batch axis when converting to tensors (see `tensor_type` above). Note that this parameter has an effect if the parameter `tensor_type` is set, *otherwise has no effect*. - **n_sequences** (`Optional[int]`, *optional*) -- You can give a tensor_type here to convert the lists of integers in PyTorch/Numpy Tensors at initialization.