Transformers documentation

OWLv2

You are viewing main version, which requires installation from source. If you'd like regular pip install, checkout the latest stable version (v4.38.1).
Hugging Face's logo
Join the Hugging Face community

and get access to the augmented documentation experience

to get started

OWLv2

Overview

OWLv2 was proposed in Scaling Open-Vocabulary Object Detection by Matthias Minderer, Alexey Gritsenko, Neil Houlsby. OWLv2 scales up OWL-ViT using self-training, which uses an existing detector to generate pseudo-box annotations on image-text pairs. This results in large gains over the previous state-of-the-art for zero-shot object detection.

The abstract from the paper is the following:

Open-vocabulary object detection has benefited greatly from pretrained vision-language models, but is still limited by the amount of available detection training data. While detection training data can be expanded by using Web image-text pairs as weak supervision, this has not been done at scales comparable to image-level pretraining. Here, we scale up detection data with self-training, which uses an existing detector to generate pseudo-box annotations on image-text pairs. Major challenges in scaling self-training are the choice of label space, pseudo-annotation filtering, and training efficiency. We present the OWLv2 model and OWL-ST self-training recipe, which address these challenges. OWLv2 surpasses the performance of previous state-of-the-art open-vocabulary detectors already at comparable training scales (~10M examples). However, with OWL-ST, we can scale to over 1B examples, yielding further large improvement: With an L/14 architecture, OWL-ST improves AP on LVIS rare classes, for which the model has seen no human box annotations, from 31.2% to 44.6% (43% relative improvement). OWL-ST unlocks Web-scale training for open-world localization, similar to what has been seen for image classification and language modelling.

drawing OWLv2 high-level overview. Taken from the original paper.

This model was contributed by nielsr. The original code can be found here.

Usage example

OWLv2 is, just like its predecessor OWL-ViT, a zero-shot text-conditioned object detection model. OWL-ViT uses CLIP as its multi-modal backbone, with a ViT-like Transformer to get visual features and a causal language model to get the text features. To use CLIP for detection, OWL-ViT removes the final token pooling layer of the vision model and attaches a lightweight classification and box head to each transformer output token. Open-vocabulary classification is enabled by replacing the fixed classification layer weights with the class-name embeddings obtained from the text model. The authors first train CLIP from scratch and fine-tune it end-to-end with the classification and box heads on standard detection datasets using a bipartite matching loss. One or multiple text queries per image can be used to perform zero-shot text-conditioned object detection.

Owlv2ImageProcessor can be used to resize (or rescale) and normalize images for the model and CLIPTokenizer is used to encode the text. Owlv2Processor wraps Owlv2ImageProcessor and CLIPTokenizer into a single instance to both encode the text and prepare the images. The following example shows how to perform object detection using Owlv2Processor and Owlv2ForObjectDetection.

>>> import requests
>>> from PIL import Image
>>> import torch

>>> from transformers import Owlv2Processor, Owlv2ForObjectDetection

>>> processor = Owlv2Processor.from_pretrained("google/owlv2-base-patch16-ensemble")
>>> model = Owlv2ForObjectDetection.from_pretrained("google/owlv2-base-patch16-ensemble")

>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)
>>> texts = [["a photo of a cat", "a photo of a dog"]]
>>> inputs = processor(text=texts, images=image, return_tensors="pt")
>>> outputs = model(**inputs)

>>> # Target image sizes (height, width) to rescale box predictions [batch_size, 2]
>>> target_sizes = torch.Tensor([image.size[::-1]])
>>> # Convert outputs (bounding boxes and class logits) to Pascal VOC Format (xmin, ymin, xmax, ymax)
>>> results = processor.post_process_object_detection(outputs=outputs, target_sizes=target_sizes, threshold=0.1)
>>> i = 0  # Retrieve predictions for the first image for the corresponding text queries
>>> text = texts[i]
>>> boxes, scores, labels = results[i]["boxes"], results[i]["scores"], results[i]["labels"]
>>> for box, score, label in zip(boxes, scores, labels):
...     box = [round(i, 2) for i in box.tolist()]
...     print(f"Detected {text[label]} with confidence {round(score.item(), 3)} at location {box}")
Detected a photo of a cat with confidence 0.614 at location [341.67, 17.54, 642.32, 278.51]
Detected a photo of a cat with confidence 0.665 at location [6.75, 38.97, 326.62, 354.85]

Resources

The architecture of OWLv2 is identical to OWL-ViT, however the object detection head now also includes an objectness classifier, which predicts the (query-agnostic) likelihood that a predicted box contains an object (as opposed to background). The objectness score can be used to rank or filter predictions independently of text queries. Usage of OWLv2 is identical to OWL-ViT with a new, updated image processor (Owlv2ImageProcessor).

Owlv2Config

class transformers.Owlv2Config

< >

( text_config = None vision_config = None projection_dim = 512 logit_scale_init_value = 2.6592 return_dict = True **kwargs )

Parameters

  • text_config (dict, optional) — Dictionary of configuration options used to initialize Owlv2TextConfig.
  • vision_config (dict, optional) — Dictionary of configuration options used to initialize Owlv2VisionConfig.
  • projection_dim (int, optional, defaults to 512) — Dimensionality of text and vision projection layers.
  • logit_scale_init_value (float, optional, defaults to 2.6592) — The inital value of the logit_scale parameter. Default is used as per the original OWLv2 implementation.
  • return_dict (bool, optional, defaults to True) — Whether or not the model should return a dictionary. If False, returns a tuple.
  • kwargs (optional) — Dictionary of keyword arguments.

Owlv2Config is the configuration class to store the configuration of an Owlv2Model. It is used to instantiate an OWLv2 model according to the specified arguments, defining the text model and vision model configs. Instantiating a configuration with the defaults will yield a similar configuration to that of the OWLv2 google/owlv2-base-patch16 architecture.

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

from_text_vision_configs

< >

( text_config: Dict vision_config: Dict **kwargs ) → Owlv2Config

Returns

Owlv2Config

An instance of a configuration object

Instantiate a Owlv2Config (or a derived class) from owlv2 text model configuration and owlv2 vision model configuration.

Owlv2TextConfig

class transformers.Owlv2TextConfig

< >

( vocab_size = 49408 hidden_size = 512 intermediate_size = 2048 num_hidden_layers = 12 num_attention_heads = 8 max_position_embeddings = 16 hidden_act = 'quick_gelu' layer_norm_eps = 1e-05 attention_dropout = 0.0 initializer_range = 0.02 initializer_factor = 1.0 pad_token_id = 0 bos_token_id = 49406 eos_token_id = 49407 **kwargs )

Parameters

  • vocab_size (int, optional, defaults to 49408) — Vocabulary size of the OWLv2 text model. Defines the number of different tokens that can be represented by the inputs_ids passed when calling Owlv2TextModel.
  • hidden_size (int, optional, defaults to 512) — Dimensionality of the encoder layers and the pooler layer.
  • intermediate_size (int, optional, defaults to 2048) — 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 8) — Number of attention heads for each attention layer in the Transformer encoder.
  • max_position_embeddings (int, optional, defaults to 16) — The maximum sequence length that this model might ever be used with. Typically set this to something large just in case (e.g., 512 or 1024 or 2048).
  • hidden_act (str or function, optional, defaults to "quick_gelu") — 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-05) — 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.
  • initializer_factor (float, optional, defaults to 1.0) — A factor for initializing all weight matrices (should be kept to 1, used internally for initialization testing).
  • pad_token_id (int, optional, defaults to 0) — The id of the padding token in the input sequences.
  • bos_token_id (int, optional, defaults to 49406) — The id of the beginning-of-sequence token in the input sequences.
  • eos_token_id (int, optional, defaults to 49407) — The id of the end-of-sequence token in the input sequences.

This is the configuration class to store the configuration of an Owlv2TextModel. It is used to instantiate an Owlv2 text 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 Owlv2 google/owlv2-base-patch16 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 Owlv2TextConfig, Owlv2TextModel

>>> # Initializing a Owlv2TextModel with google/owlv2-base-patch16 style configuration
>>> configuration = Owlv2TextConfig()

>>> # Initializing a Owlv2TextConfig from the google/owlv2-base-patch16 style configuration
>>> model = Owlv2TextModel(configuration)

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

Owlv2VisionConfig

class transformers.Owlv2VisionConfig

< >

( hidden_size = 768 intermediate_size = 3072 num_hidden_layers = 12 num_attention_heads = 12 num_channels = 3 image_size = 768 patch_size = 16 hidden_act = 'quick_gelu' layer_norm_eps = 1e-05 attention_dropout = 0.0 initializer_range = 0.02 initializer_factor = 1.0 **kwargs )

Parameters

  • hidden_size (int, optional, defaults to 768) — 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 12) — 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 768) — The size (resolution) of each image.
  • patch_size (int, optional, defaults to 16) — The size (resolution) of each patch.
  • hidden_act (str or function, optional, defaults to "quick_gelu") — 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-05) — 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.
  • initializer_factor (float, optional, defaults to 1.0) — A factor for initializing all weight matrices (should be kept to 1, used internally for initialization testing).

This is the configuration class to store the configuration of an Owlv2VisionModel. It is used to instantiate an OWLv2 image 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 OWLv2 google/owlv2-base-patch16 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 Owlv2VisionConfig, Owlv2VisionModel

>>> # Initializing a Owlv2VisionModel with google/owlv2-base-patch16 style configuration
>>> configuration = Owlv2VisionConfig()

>>> # Initializing a Owlv2VisionModel model from the google/owlv2-base-patch16 style configuration
>>> model = Owlv2VisionModel(configuration)

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

Owlv2ImageProcessor

class transformers.Owlv2ImageProcessor

< >

( do_rescale: bool = True rescale_factor: Union = 0.00392156862745098 do_pad: bool = True do_resize: bool = True size: Dict = None resample: Resampling = <Resampling.BILINEAR: 2> do_normalize: bool = True image_mean: Union = None image_std: Union = None **kwargs )

Parameters

  • do_rescale (bool, optional, defaults to True) — Whether to rescale the image by the specified scale rescale_factor. Can be overriden by do_rescale in the preprocess method.
  • rescale_factor (int or float, optional, defaults to 1/255) — Scale factor to use if rescaling the image. Can be overriden by rescale_factor in the preprocess method.
  • do_pad (bool, optional, defaults to True) — Whether to pad the image to a square with gray pixels on the bottom and the right. Can be overriden by do_pad in the preprocess method.
  • do_resize (bool, optional, defaults to True) — Controls whether to resize the image’s (height, width) dimensions to the specified size. Can be overriden by do_resize in the preprocess method.
  • size (Dict[str, int] optional, defaults to {"height" -- 960, "width": 960}): Size to resize the image to. Can be overriden by size in the preprocess method.
  • resample (PILImageResampling, optional, defaults to Resampling.BILINEAR) — Resampling method to use if resizing the image. Can be overriden by resample in the preprocess method.
  • do_normalize (bool, optional, defaults to True) — Whether to normalize the image. Can be overridden by the do_normalize parameter in the preprocess method.
  • image_mean (float or List[float], optional, defaults to OPENAI_CLIP_MEAN) — Mean to use if normalizing the image. This is a float or list of floats the length of the number of channels in the image. Can be overridden by the image_mean parameter in the preprocess method.
  • image_std (float or List[float], optional, defaults to OPENAI_CLIP_STD) — Standard deviation to use if normalizing the image. This is a float or list of floats the length of the number of channels in the image. Can be overridden by the image_std parameter in the preprocess method.

Constructs an OWLv2 image processor.

preprocess

< >

( images: Union do_pad: bool = None do_resize: bool = None size: Dict = None do_rescale: bool = None rescale_factor: float = None do_normalize: bool = None image_mean: Union = None image_std: Union = None return_tensors: Union = None data_format: ChannelDimension = <ChannelDimension.FIRST: 'channels_first'> input_data_format: Union = None **kwargs )

Parameters

  • images (ImageInput) — Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If passing in images with pixel values between 0 and 1, set do_rescale=False.
  • do_pad (bool, optional, defaults to self.do_pad) — Whether to pad the image to a square with gray pixels on the bottom and the right.
  • do_resize (bool, optional, defaults to self.do_resize) — Whether to resize the image.
  • size (Dict[str, int], optional, defaults to self.size) — Size to resize the image to.
  • do_rescale (bool, optional, defaults to self.do_rescale) — Whether to rescale the image values between [0 - 1].
  • 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.
  • image_std (float or List[float], optional, defaults to self.image_std) — Image standard deviation.
  • return_tensors (str or TensorType, optional) — The type of tensors to return. Can be one of:
    • Unset: Return a list of np.ndarray.
    • TensorType.TENSORFLOW or 'tf': Return a batch of type tf.Tensor.
    • TensorType.PYTORCH or 'pt': Return a batch of type torch.Tensor.
    • TensorType.NUMPY or 'np': Return a batch of type np.ndarray.
    • TensorType.JAX or 'jax': Return a batch of type jax.numpy.ndarray.
  • 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.

Preprocess an image or batch of images.

post_process_object_detection

< >

( outputs threshold: float = 0.1 target_sizes: Union = None ) → List[Dict]

Parameters

  • outputs (OwlViTObjectDetectionOutput) — Raw outputs of the model.
  • threshold (float, optional) — Score threshold to keep object detection predictions.
  • target_sizes (torch.Tensor or List[Tuple[int, int]], optional) — Tensor of shape (batch_size, 2) or list of tuples (Tuple[int, int]) containing the target size (height, width) of each image in the batch. If unset, predictions will not be resized.

Returns

List[Dict]

A list of dictionaries, each dictionary containing the scores, labels and boxes for an image in the batch as predicted by the model.

Converts the raw output of OwlViTForObjectDetection into final bounding boxes in (top_left_x, top_left_y, bottom_right_x, bottom_right_y) format.

post_process_image_guided_detection

< >

( outputs threshold = 0.0 nms_threshold = 0.3 target_sizes = None ) → List[Dict]

Parameters

  • outputs (OwlViTImageGuidedObjectDetectionOutput) — Raw outputs of the model.
  • threshold (float, optional, defaults to 0.0) — Minimum confidence threshold to use to filter out predicted boxes.
  • nms_threshold (float, optional, defaults to 0.3) — IoU threshold for non-maximum suppression of overlapping boxes.
  • target_sizes (torch.Tensor, optional) — Tensor of shape (batch_size, 2) where each entry is the (height, width) of the corresponding image in the batch. If set, predicted normalized bounding boxes are rescaled to the target sizes. If left to None, predictions will not be unnormalized.

Returns

List[Dict]

A list of dictionaries, each dictionary containing the scores, labels and boxes for an image in the batch as predicted by the model. All labels are set to None as OwlViTForObjectDetection.image_guided_detection perform one-shot object detection.

Converts the output of OwlViTForObjectDetection.image_guided_detection() into the format expected by the COCO api.

Owlv2Processor

class transformers.Owlv2Processor

< >

( image_processor tokenizer **kwargs )

Parameters

  • image_processor (Owlv2ImageProcessor) — The image processor is a required input.
  • tokenizer ([CLIPTokenizer, CLIPTokenizerFast]) — The tokenizer is a required input.

Constructs an Owlv2 processor which wraps Owlv2ImageProcessor and CLIPTokenizer/CLIPTokenizerFast into a single processor that interits both the image processor and tokenizer functionalities. See the __call__() and decode() for more information.

batch_decode

< >

( *args **kwargs )

This method forwards all its arguments to CLIPTokenizerFast’s batch_decode(). Please refer to the docstring of this method for more information.

decode

< >

( *args **kwargs )

This method forwards all its arguments to CLIPTokenizerFast’s decode(). Please refer to the docstring of this method for more information.

post_process_image_guided_detection

< >

( *args **kwargs )

This method forwards all its arguments to OwlViTImageProcessor.post_process_one_shot_object_detection. Please refer to the docstring of this method for more information.

post_process_object_detection

< >

( *args **kwargs )

This method forwards all its arguments to OwlViTImageProcessor.post_process_object_detection(). Please refer to the docstring of this method for more information.

Owlv2Model

class transformers.Owlv2Model

< >

( config: Owlv2Config )

Parameters

  • config (Owvl2Config) — 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.

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

< >

( input_ids: Optional = None pixel_values: Optional = None attention_mask: Optional = None return_loss: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_base_image_embeds: Optional = None return_dict: Optional = None ) → transformers.models.owlv2.modeling_owlv2.Owlv2Output or tuple(torch.FloatTensor)

Parameters

  • input_ids (torch.LongTensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details. What are input IDs?
  • pixel_values (torch.FloatTensor of shape (batch_size, num_channels, height, width)) — Pixel values.
  • 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]:
  • return_loss (bool, optional) — Whether or not to return the contrastive loss.
  • 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.
  • return_base_image_embeds (bool, optional) — Whether or not to return the base image embeddings.
  • return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

Returns

transformers.models.owlv2.modeling_owlv2.Owlv2Output or tuple(torch.FloatTensor)

A transformers.models.owlv2.modeling_owlv2.Owlv2Output 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 (<class 'transformers.models.owlv2.configuration_owlv2.Owlv2Config'>) and inputs.

  • loss (torch.FloatTensor of shape (1,), optional, returned when return_loss is True) — Contrastive loss for image-text similarity.
  • logits_per_image (torch.FloatTensor of shape (image_batch_size, text_batch_size)) — The scaled dot product scores between image_embeds and text_embeds. This represents the image-text similarity scores.
  • logits_per_text (torch.FloatTensor of shape (text_batch_size, image_batch_size)) — The scaled dot product scores between text_embeds and image_embeds. This represents the text-image similarity scores.
  • text_embeds (torch.FloatTensor of shape (batch_size * num_max_text_queries, output_dim) — The text embeddings obtained by applying the projection layer to the pooled output of Owlv2TextModel.
  • image_embeds (torch.FloatTensor of shape (batch_size, output_dim) — The image embeddings obtained by applying the projection layer to the pooled output of Owlv2VisionModel.
  • text_model_output (TupleBaseModelOutputWithPooling) — The output of the Owlv2TextModel.
  • vision_model_output (BaseModelOutputWithPooling) — The output of the Owlv2VisionModel.

The Owlv2Model 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.

Examples:

>>> from PIL import Image
>>> import requests
>>> from transformers import AutoProcessor, Owlv2Model

>>> model = Owlv2Model.from_pretrained("google/owlv2-base-patch16-ensemble")
>>> processor = AutoProcessor.from_pretrained("google/owlv2-base-patch16-ensemble")
>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)
>>> inputs = processor(text=[["a photo of a cat", "a photo of a dog"]], images=image, return_tensors="pt")
>>> outputs = model(**inputs)
>>> logits_per_image = outputs.logits_per_image  # this is the image-text similarity score
>>> probs = logits_per_image.softmax(dim=1)  # we can take the softmax to get the label probabilities

get_text_features

< >

( input_ids: Optional = None attention_mask: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → text_features (torch.FloatTensor of shape (batch_size, output_dim)

Parameters

  • input_ids (torch.LongTensor of shape (batch_size * num_max_text_queries, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details. What are input IDs?
  • attention_mask (torch.Tensor of shape (batch_size, num_max_text_queries, sequence_length), optional) — Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
  • 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.
  • return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

Returns

text_features (torch.FloatTensor of shape (batch_size, output_dim)

The text embeddings obtained by applying the projection layer to the pooled output of Owlv2TextModel.

The Owlv2Model 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.

Examples:

>>> from transformers import AutoProcessor, Owlv2Model

>>> model = Owlv2Model.from_pretrained("google/owlv2-base-patch16-ensemble")
>>> processor = AutoProcessor.from_pretrained("google/owlv2-base-patch16-ensemble")
>>> inputs = processor(
...     text=[["a photo of a cat", "a photo of a dog"], ["photo of a astranaut"]], return_tensors="pt"
... )
>>> text_features = model.get_text_features(**inputs)

get_image_features

< >

( pixel_values: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → image_features (torch.FloatTensor of shape (batch_size, output_dim)

Parameters

  • pixel_values (torch.FloatTensor of shape (batch_size, num_channels, height, width)) — Pixel 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.
  • return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

Returns

image_features (torch.FloatTensor of shape (batch_size, output_dim)

The image embeddings obtained by applying the projection layer to the pooled output of Owlv2VisionModel.

The Owlv2Model 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.

Examples:

>>> from PIL import Image
>>> import requests
>>> from transformers import AutoProcessor, Owlv2Model

>>> model = Owlv2Model.from_pretrained("google/owlv2-base-patch16-ensemble")
>>> processor = AutoProcessor.from_pretrained("google/owlv2-base-patch16-ensemble")
>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)
>>> inputs = processor(images=image, return_tensors="pt")
>>> image_features = model.get_image_features(**inputs)

Owlv2TextModel

class transformers.Owlv2TextModel

< >

( config: Owlv2TextConfig )

forward

< >

( input_ids: Tensor attention_mask: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.modeling_outputs.BaseModelOutputWithPooling or tuple(torch.FloatTensor)

Parameters

  • input_ids (torch.LongTensor of shape (batch_size * num_max_text_queries, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details. What are input IDs?
  • attention_mask (torch.Tensor of shape (batch_size, num_max_text_queries, sequence_length), optional) — Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
  • 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.
  • 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 (<class 'transformers.models.owlv2.configuration_owlv2.Owlv2TextConfig'>) 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 Owlv2TextModel 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.

Examples:

>>> from transformers import AutoProcessor, Owlv2TextModel

>>> model = Owlv2TextModel.from_pretrained("google/owlv2-base-patch16")
>>> processor = AutoProcessor.from_pretrained("google/owlv2-base-patch16")
>>> inputs = processor(
...     text=[["a photo of a cat", "a photo of a dog"], ["photo of a astranaut"]], return_tensors="pt"
... )
>>> outputs = model(**inputs)
>>> last_hidden_state = outputs.last_hidden_state
>>> pooled_output = outputs.pooler_output  # pooled (EOS token) states

Owlv2VisionModel

class transformers.Owlv2VisionModel

< >

( config: Owlv2VisionConfig )

forward

< >

( pixel_values: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.modeling_outputs.BaseModelOutputWithPooling or tuple(torch.FloatTensor)

Parameters

  • pixel_values (torch.FloatTensor of shape (batch_size, num_channels, height, width)) — Pixel 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.
  • 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 (<class 'transformers.models.owlv2.configuration_owlv2.Owlv2VisionConfig'>) 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 Owlv2VisionModel 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.

Examples:

>>> from PIL import Image
>>> import requests
>>> from transformers import AutoProcessor, Owlv2VisionModel

>>> model = Owlv2VisionModel.from_pretrained("google/owlv2-base-patch16")
>>> processor = AutoProcessor.from_pretrained("google/owlv2-base-patch16")
>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)

>>> inputs = processor(images=image, return_tensors="pt")

>>> outputs = model(**inputs)
>>> last_hidden_state = outputs.last_hidden_state
>>> pooled_output = outputs.pooler_output  # pooled CLS states

Owlv2ForObjectDetection

class transformers.Owlv2ForObjectDetection

< >

( config: Owlv2Config )

forward

< >

( input_ids: Tensor pixel_values: FloatTensor attention_mask: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.models.owlv2.modeling_owlv2.Owlv2ObjectDetectionOutput or tuple(torch.FloatTensor)

Parameters

  • pixel_values (torch.FloatTensor of shape (batch_size, num_channels, height, width)) — Pixel values.
  • input_ids (torch.LongTensor of shape (batch_size * num_max_text_queries, sequence_length), optional) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details. What are input IDs?.
  • attention_mask (torch.Tensor of shape (batch_size, num_max_text_queries, sequence_length), optional) — Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
  • output_hidden_states (bool, optional) — Whether or not to return the last hidden state. See text_model_last_hidden_state and vision_model_last_hidden_state under returned tensors for more detail.
  • return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

Returns

transformers.models.owlv2.modeling_owlv2.Owlv2ObjectDetectionOutput or tuple(torch.FloatTensor)

A transformers.models.owlv2.modeling_owlv2.Owlv2ObjectDetectionOutput 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 (<class 'transformers.models.owlv2.configuration_owlv2.Owlv2Config'>) and inputs.

  • loss (torch.FloatTensor of shape (1,), optional, returned when labels are provided)) — Total loss as a linear combination of a negative log-likehood (cross-entropy) for class prediction and a bounding box loss. The latter is defined as a linear combination of the L1 loss and the generalized scale-invariant IoU loss.
  • loss_dict (Dict, optional) — A dictionary containing the individual losses. Useful for logging.
  • logits (torch.FloatTensor of shape (batch_size, num_patches, num_queries)) — Classification logits (including no-object) for all queries.
  • objectness_logits (torch.FloatTensor of shape (batch_size, num_patches, 1)) — The objectness logits of all image patches. OWL-ViT represents images as a set of image patches where the total number of patches is (image_size / patch_size)**2.
  • pred_boxes (torch.FloatTensor of shape (batch_size, num_patches, 4)) — Normalized boxes coordinates for all queries, represented as (center_x, center_y, width, height). These values are normalized in [0, 1], relative to the size of each individual image in the batch (disregarding possible padding). You can use post_process_object_detection() to retrieve the unnormalized bounding boxes.
  • text_embeds (torch.FloatTensor of shape (batch_size, num_max_text_queries, output_dim) — The text embeddings obtained by applying the projection layer to the pooled output of Owlv2TextModel.
  • image_embeds (torch.FloatTensor of shape (batch_size, patch_size, patch_size, output_dim) — Pooled output of Owlv2VisionModel. OWLv2 represents images as a set of image patches and computes image embeddings for each patch.
  • class_embeds (torch.FloatTensor of shape (batch_size, num_patches, hidden_size)) — Class embeddings of all image patches. OWLv2 represents images as a set of image patches where the total number of patches is (image_size / patch_size)**2.
  • text_model_output (TupleBaseModelOutputWithPooling) — The output of the Owlv2TextModel.
  • vision_model_output (BaseModelOutputWithPooling) — The output of the Owlv2VisionModel.

The Owlv2ForObjectDetection 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.

Examples:

>>> import requests
>>> from PIL import Image
>>> import numpy as np
>>> import torch
>>> from transformers import AutoProcessor, Owlv2ForObjectDetection
>>> from transformers.utils.constants import OPENAI_CLIP_MEAN, OPENAI_CLIP_STD

>>> processor = AutoProcessor.from_pretrained("google/owlv2-base-patch16-ensemble")
>>> model = Owlv2ForObjectDetection.from_pretrained("google/owlv2-base-patch16-ensemble")

>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)
>>> texts = [["a photo of a cat", "a photo of a dog"]]
>>> inputs = processor(text=texts, images=image, return_tensors="pt")

>>> # forward pass
>>> with torch.no_grad():
...     outputs = model(**inputs)

>>> # Note: boxes need to be visualized on the padded, unnormalized image
>>> # hence we'll set the target image sizes (height, width) based on that

>>> def get_preprocessed_image(pixel_values):
...     pixel_values = pixel_values.squeeze().numpy()
...     unnormalized_image = (pixel_values * np.array(OPENAI_CLIP_STD)[:, None, None]) + np.array(OPENAI_CLIP_MEAN)[:, None, None]
...     unnormalized_image = (unnormalized_image * 255).astype(np.uint8)
...     unnormalized_image = np.moveaxis(unnormalized_image, 0, -1)
...     unnormalized_image = Image.fromarray(unnormalized_image)
...     return unnormalized_image

>>> unnormalized_image = get_preprocessed_image(inputs.pixel_values)

>>> target_sizes = torch.Tensor([unnormalized_image.size[::-1]])
>>> # Convert outputs (bounding boxes and class logits) to final bounding boxes and scores
>>> results = processor.post_process_object_detection(
...     outputs=outputs, threshold=0.2, target_sizes=target_sizes
... )

>>> i = 0  # Retrieve predictions for the first image for the corresponding text queries
>>> text = texts[i]
>>> boxes, scores, labels = results[i]["boxes"], results[i]["scores"], results[i]["labels"]

>>> for box, score, label in zip(boxes, scores, labels):
...     box = [round(i, 2) for i in box.tolist()]
...     print(f"Detected {text[label]} with confidence {round(score.item(), 3)} at location {box}")
Detected a photo of a cat with confidence 0.614 at location [512.5, 35.08, 963.48, 557.02]
Detected a photo of a cat with confidence 0.665 at location [10.13, 77.94, 489.93, 709.69]

image_guided_detection

< >

( pixel_values: FloatTensor query_pixel_values: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.models.owlv2.modeling_owlv2.Owlv2ImageGuidedObjectDetectionOutput or tuple(torch.FloatTensor)

Parameters

  • pixel_values (torch.FloatTensor of shape (batch_size, num_channels, height, width)) — Pixel values.
  • query_pixel_values (torch.FloatTensor of shape (batch_size, num_channels, height, width)) — Pixel values of query image(s) to be detected. Pass in one query image per target image.
  • 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.
  • return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.

Returns

transformers.models.owlv2.modeling_owlv2.Owlv2ImageGuidedObjectDetectionOutput or tuple(torch.FloatTensor)

A transformers.models.owlv2.modeling_owlv2.Owlv2ImageGuidedObjectDetectionOutput 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 (<class 'transformers.models.owlv2.configuration_owlv2.Owlv2Config'>) and inputs.

  • logits (torch.FloatTensor of shape (batch_size, num_patches, num_queries)) — Classification logits (including no-object) for all queries.
  • target_pred_boxes (torch.FloatTensor of shape (batch_size, num_patches, 4)) — Normalized boxes coordinates for all queries, represented as (center_x, center_y, width, height). These values are normalized in [0, 1], relative to the size of each individual target image in the batch (disregarding possible padding). You can use post_process_object_detection() to retrieve the unnormalized bounding boxes.
  • query_pred_boxes (torch.FloatTensor of shape (batch_size, num_patches, 4)) — Normalized boxes coordinates for all queries, represented as (center_x, center_y, width, height). These values are normalized in [0, 1], relative to the size of each individual query image in the batch (disregarding possible padding). You can use post_process_object_detection() to retrieve the unnormalized bounding boxes.
  • image_embeds (torch.FloatTensor of shape (batch_size, patch_size, patch_size, output_dim) — Pooled output of Owlv2VisionModel. OWLv2 represents images as a set of image patches and computes image embeddings for each patch.
  • query_image_embeds (torch.FloatTensor of shape (batch_size, patch_size, patch_size, output_dim) — Pooled output of Owlv2VisionModel. OWLv2 represents images as a set of image patches and computes image embeddings for each patch.
  • class_embeds (torch.FloatTensor of shape (batch_size, num_patches, hidden_size)) — Class embeddings of all image patches. OWLv2 represents images as a set of image patches where the total number of patches is (image_size / patch_size)**2.
  • text_model_output (TupleBaseModelOutputWithPooling) — The output of the Owlv2TextModel.
  • vision_model_output (BaseModelOutputWithPooling) — The output of the Owlv2VisionModel.

The Owlv2ForObjectDetection 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.

Examples:

>>> import requests
>>> from PIL import Image
>>> import torch
>>> import numpy as np
>>> from transformers import AutoProcessor, Owlv2ForObjectDetection
>>> from transformers.utils.constants import OPENAI_CLIP_MEAN, OPENAI_CLIP_STD

>>> processor = AutoProcessor.from_pretrained("google/owlv2-base-patch16-ensemble")
>>> model = Owlv2ForObjectDetection.from_pretrained("google/owlv2-base-patch16-ensemble")

>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)
>>> query_url = "http://images.cocodataset.org/val2017/000000001675.jpg"
>>> query_image = Image.open(requests.get(query_url, stream=True).raw)
>>> inputs = processor(images=image, query_images=query_image, return_tensors="pt")

>>> # forward pass
>>> with torch.no_grad():
...     outputs = model.image_guided_detection(**inputs)

>>> # Note: boxes need to be visualized on the padded, unnormalized image
>>> # hence we'll set the target image sizes (height, width) based on that

>>> def get_preprocessed_image(pixel_values):
...     pixel_values = pixel_values.squeeze().numpy()
...     unnormalized_image = (pixel_values * np.array(OPENAI_CLIP_STD)[:, None, None]) + np.array(OPENAI_CLIP_MEAN)[:, None, None]
...     unnormalized_image = (unnormalized_image * 255).astype(np.uint8)
...     unnormalized_image = np.moveaxis(unnormalized_image, 0, -1)
...     unnormalized_image = Image.fromarray(unnormalized_image)
...     return unnormalized_image

>>> unnormalized_image = get_preprocessed_image(inputs.pixel_values)

>>> target_sizes = torch.Tensor([unnormalized_image.size[::-1]])

>>> # Convert outputs (bounding boxes and class logits) to Pascal VOC format (xmin, ymin, xmax, ymax)
>>> results = processor.post_process_image_guided_detection(
...     outputs=outputs, threshold=0.9, nms_threshold=0.3, target_sizes=target_sizes
... )
>>> i = 0  # Retrieve predictions for the first image
>>> boxes, scores = results[i]["boxes"], results[i]["scores"]
>>> for box, score in zip(boxes, scores):
...     box = [round(i, 2) for i in box.tolist()]
...     print(f"Detected similar object with confidence {round(score.item(), 3)} at location {box}")
Detected similar object with confidence 0.938 at location [490.96, 109.89, 821.09, 536.11]
Detected similar object with confidence 0.959 at location [8.67, 721.29, 928.68, 732.78]
Detected similar object with confidence 0.902 at location [4.27, 720.02, 941.45, 761.59]
Detected similar object with confidence 0.985 at location [265.46, -58.9, 1009.04, 365.66]
Detected similar object with confidence 1.0 at location [9.79, 28.69, 937.31, 941.64]
Detected similar object with confidence 0.998 at location [869.97, 58.28, 923.23, 978.1]
Detected similar object with confidence 0.985 at location [309.23, 21.07, 371.61, 932.02]
Detected similar object with confidence 0.947 at location [27.93, 859.45, 969.75, 915.44]
Detected similar object with confidence 0.996 at location [785.82, 41.38, 880.26, 966.37]
Detected similar object with confidence 0.998 at location [5.08, 721.17, 925.93, 998.41]
Detected similar object with confidence 0.969 at location [6.7, 898.1, 921.75, 949.51]
Detected similar object with confidence 0.966 at location [47.16, 927.29, 981.99, 942.14]
Detected similar object with confidence 0.924 at location [46.4, 936.13, 953.02, 950.78]