I-JEPA

I-JEPA is a self-supervised learning method that learns semantic image representations by predicting parts of an image from other parts of the image. It compares the abstract representations of the image (rather than pixel level comparisons), which avoids the typical pitfalls of data augmentation bias and pixel-level details that don’t capture semantic meaning.

You can find the original I-JEPA checkpoints under the AI at Meta organization.

This model was contributed by jmtzt.

Click on the I-JEPA models in the right sidebar for more examples of how to apply I-JEPA to different image representation and classification tasks.

The example below demonstrates how to extract image features with Pipeline or the AutoModel class.

Pipeline

AutoModel

Quantization reduces the memory burden of large models by representing the weights in a lower precision. Refer to the Quantization overview for more available quantization backends. The example below uses bitsandbytes to only quantize the weights to 4-bits.

import torch
from transformers import BitsAndBytesConfig, AutoModel, AutoProcessor
from datasets import load_dataset

quantization_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_compute_dtype=torch.bfloat16,
    bnb_4bit_use_double_quant=True,
)

url_1 = "http://images.cocodataset.org/val2017/000000039769.jpg"
url_2 = "http://images.cocodataset.org/val2017/000000219578.jpg"
image_1 = Image.open(requests.get(url_1, stream=True).raw)
image_2 = Image.open(requests.get(url_2, stream=True).raw)

processor = AutoProcessor.from_pretrained("facebook/ijepa_vitg16_22k")
model = AutoModel.from_pretrained("facebook/ijepa_vitg16_22k", quantization_config=quantization_config, torch_dtype="auto", attn_implementation="sdpa")


def infer(image):
    inputs = processor(image, return_tensors="pt")
    outputs = model(**inputs)
    return outputs.last_hidden_state.mean(dim=1)


embed_1 = infer(image_1)
embed_2 = infer(image_2)

similarity = cosine_similarity(embed_1, embed_2)
print(similarity)

IJepaConfig

class transformers.IJepaConfig

< source >

( hidden_size = 768 num_hidden_layers = 12 num_attention_heads = 12 intermediate_size = 3072 hidden_act = 'gelu' hidden_dropout_prob = 0.0 attention_probs_dropout_prob = 0.0 initializer_range = 0.02 layer_norm_eps = 1e-12 image_size = 224 patch_size = 16 num_channels = 3 qkv_bias = True pooler_output_size = None pooler_act = 'tanh' **kwargs )

Parameters

hidden_size (int, optional, defaults to 768) — Dimensionality of the encoder layers and the pooler layer.
num_hidden_layers (int, optional, defaults to 12) — Number of hidden layers in the Transformer encoder.
num_attention_heads (int, optional, defaults to 12) — Number of attention heads for each attention layer in the Transformer encoder.
intermediate_size (int, optional, defaults to 3072) — Dimensionality of the “intermediate” (i.e., feed-forward) layer in the Transformer encoder.
hidden_act (str or function, optional, defaults to "gelu") — The non-linear activation function (function or string) in the encoder and pooler. If string, "gelu", "relu", "selu" and "gelu_new" are supported.
hidden_dropout_prob (float, optional, defaults to 0.0) — The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.
attention_probs_dropout_prob (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.
layer_norm_eps (float, optional, defaults to 1e-12) — The epsilon used by the layer normalization layers.
image_size (int, optional, defaults to 224) — The size (resolution) of each image.
patch_size (int, optional, defaults to 16) — The size (resolution) of each patch.
num_channels (int, optional, defaults to 3) — The number of input channels.
qkv_bias (bool, optional, defaults to True) — Whether to add a bias to the queries, keys and values.
pooler_output_size (int, optional) — Dimensionality of the pooler layer. If None, defaults to hidden_size.
pooler_act (str, optional, defaults to "tanh") — The activation function to be used by the pooler. Keys of ACT2FN are supported for Flax and Pytorch, and elements of https://www.tensorflow.org/api_docs/python/tf/keras/activations are supported for Tensorflow.

This is the configuration class to store the configuration of a IJepaModel. It is used to instantiate an IJEPA 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 I-JEPA facebook/ijepa_vith14_1k architecture.

Configuration objects inherit from PretrainedConfig and can be used to control the model outputs. Read the documentation from PretrainedConfig for more information.

Example:

>>> from transformers import IJepaConfig, IJepaModel

>>> # Initializing a IJEPA ijepa-base-patch16-224 style configuration
>>> configuration = IJepaConfig()

>>> # Initializing a model (with random weights) from the ijepa-base-patch16-224 style configuration
>>> model = IJepaModel(configuration)

>>> # Accessing the model configuration
>>> configuration = model.config

IJepaModel

class transformers.IJepaModel

< source >

( config: IJepaConfig add_pooling_layer: bool = False use_mask_token: bool = False )

Parameters

config (IJepaConfig) — 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() method to load the model weights.
add_pooling_layer (bool, optional, defaults to True) — Whether to add a pooling layer
use_mask_token (bool, optional, defaults to False) — Whether to use a mask token for masked image modeling.

The bare Ijepa Model outputting raw hidden-states without any specific head on top.

This model inherits from 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 subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

forward

< source >

( pixel_values: typing.Optional[torch.Tensor] = None bool_masked_pos: typing.Optional[torch.BoolTensor] = None head_mask: typing.Optional[torch.Tensor] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None interpolate_pos_encoding: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.modeling_outputs.BaseModelOutputWithPooling or tuple(torch.FloatTensor)

Parameters

pixel_values (torch.Tensor of shape (batch_size, num_channels, image_size, image_size), optional) — The tensors corresponding to the input images. Pixel values can be obtained using {image_processor_class}. See {image_processor_class}.__call__ for details ({processor_class} uses {image_processor_class} for processing images).
bool_masked_pos (torch.BoolTensor of shape (batch_size, num_patches), optional) — Boolean masked positions. Indicates which patches are masked (1) and which aren’t (0).
head_mask (torch.Tensor of shape (num_heads,) or (num_layers, num_heads), optional) — Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:
- 1 indicates the head is not masked,
- 0 indicates the head is masked.
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.
interpolate_pos_encoding (bool, optional) — Whether to interpolate the pre-trained position encodings.
return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

Returns

transformers.modeling_outputs.BaseModelOutputWithPooling or tuple(torch.FloatTensor)

A 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 (IJepaConfig) 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.

The IJepaModel 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:

IJepaForImageClassification

class transformers.IJepaForImageClassification

< source >

( config: IJepaConfig )

Parameters

config (IJepaConfig) — 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() method to load the model weights.

IJepa Model transformer with an image classification head on top (a linear layer on top of the final hidden states) e.g. for ImageNet.

Note that it’s possible to fine-tune IJepa on higher resolution images than the ones it has been trained on, by setting interpolate_pos_encoding to True in the forward of the model. This will interpolate the pre-trained position embeddings to the higher resolution.

This model is also a PyTorch 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.

forward

< source >

( pixel_values: typing.Optional[torch.Tensor] = None head_mask: typing.Optional[torch.Tensor] = None labels: typing.Optional[torch.Tensor] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None interpolate_pos_encoding: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.modeling_outputs.ImageClassifierOutput or tuple(torch.FloatTensor)

Parameters

pixel_values (torch.Tensor of shape (batch_size, num_channels, image_size, image_size), optional) — The tensors corresponding to the input images. Pixel values can be obtained using {image_processor_class}. See {image_processor_class}.__call__ for details ({processor_class} uses {image_processor_class} for processing images).
head_mask (torch.Tensor of shape (num_heads,) or (num_layers, num_heads), optional) — Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:
- 1 indicates the head is not masked,
- 0 indicates the head is masked.
labels (torch.LongTensor of shape (batch_size,), optional) — Labels for computing the image classification/regression loss. Indices should be in [0, ..., config.num_labels - 1]. If config.num_labels == 1 a regression loss is computed (Mean-Square loss), If config.num_labels > 1 a classification loss is computed (Cross-Entropy).
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.
interpolate_pos_encoding (bool, optional) — Whether to interpolate the pre-trained position encodings.
return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

Returns

transformers.modeling_outputs.ImageClassifierOutput or tuple(torch.FloatTensor)

A transformers.modeling_outputs.ImageClassifierOutput 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 (IJepaConfig) and inputs.

loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) — Classification (or regression if config.num_labels==1) loss.
logits (torch.FloatTensor of shape (batch_size, config.num_labels)) — Classification (or regression if config.num_labels==1) scores (before SoftMax).
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 stage) of shape (batch_size, sequence_length, hidden_size). Hidden-states (also called feature maps) of the model at the output of each stage.
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, patch_size, sequence_length).

Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

The IJepaForImageClassification forward method, overrides the __call__ special method.

Example:

>>> from transformers import AutoImageProcessor, IJepaForImageClassification
>>> import torch
>>> from datasets import load_dataset

>>> dataset = load_dataset("huggingface/cats-image")
>>> image = dataset["test"]["image"][0]

>>> image_processor = AutoImageProcessor.from_pretrained("facebook/ijepa_vith14_1k")
>>> model = IJepaForImageClassification.from_pretrained("facebook/ijepa_vith14_1k")

>>> inputs = image_processor(image, return_tensors="pt")

>>> with torch.no_grad():
...     logits = model(**inputs).logits

>>> # model predicts one of the 1000 ImageNet classes
>>> predicted_label = logits.argmax(-1).item()
>>> print(model.config.id2label[predicted_label])
...

< > Update on GitHub

Transformers

I-JEPA

IJepaConfig

class transformers.IJepaConfig

IJepaModel

class transformers.IJepaModel

forward

IJepaForImageClassification

class transformers.IJepaForImageClassification

forward