The MobileViTV2 model was proposed in Separable Self-attention for Mobile Vision Transformers by Sachin Mehta and Mohammad Rastegari.
MobileViTV2 is the second version of MobileViT, constructed by replacing the multi-headed self-attention in MobileViT with separable self-attention.
The abstract from the paper is the following:
Mobile vision transformers (MobileViT) can achieve state-of-the-art performance across several mobile vision tasks, including classification and detection. Though these models have fewer parameters, they have high latency as compared to convolutional neural network-based models. The main efficiency bottleneck in MobileViT is the multi-headed self-attention (MHA) in transformers, which requires O(k2) time complexity with respect to the number of tokens (or patches) k. Moreover, MHA requires costly operations (e.g., batch-wise matrix multiplication) for computing self-attention, impacting latency on resource-constrained devices. This paper introduces a separable self-attention method with linear complexity, i.e. O(k). A simple yet effective characteristic of the proposed method is that it uses element-wise operations for computing self-attention, making it a good choice for resource-constrained devices. The improved model, MobileViTV2, is state-of-the-art on several mobile vision tasks, including ImageNet object classification and MS-COCO object detection. With about three million parameters, MobileViTV2 achieves a top-1 accuracy of 75.6% on the ImageNet dataset, outperforming MobileViT by about 1% while running 3.2× faster on a mobile device.
This model was contributed by shehan97. The original code can be found here.
( num_channels = 3 image_size = 256 patch_size = 2 expand_ratio = 2.0 hidden_act = 'swish' conv_kernel_size = 3 output_stride = 32 classifier_dropout_prob = 0.1 initializer_range = 0.02 layer_norm_eps = 1e-05 aspp_out_channels = 512 atrous_rates = [6, 12, 18] aspp_dropout_prob = 0.1 semantic_loss_ignore_index = 255 n_attn_blocks = [2, 4, 3] base_attn_unit_dims = [128, 192, 256] width_multiplier = 1.0 ffn_multiplier = 2 attn_dropout = 0.0 ffn_dropout = 0.0 **kwargs )
Parameters
int, optional, defaults to 3) —
The number of input channels. int, optional, defaults to 256) —
The size (resolution) of each image. int, optional, defaults to 2) —
The size (resolution) of each patch. float, optional, defaults to 2.0) —
Expansion factor for the MobileNetv2 layers. str or function, optional, defaults to "swish") —
The non-linear activation function (function or string) in the Transformer encoder and convolution layers. int, optional, defaults to 3) —
The size of the convolutional kernel in the MobileViTV2 layer. int, optional, defaults to 32) —
The ratio of the spatial resolution of the output to the resolution of the input image. float, optional, defaults to 0.1) —
The dropout ratio for attached classifiers. float, optional, defaults to 0.02) —
The standard deviation of the truncated_normal_initializer for initializing all weight matrices. float, optional, defaults to 1e-05) —
The epsilon used by the layer normalization layers. int, optional, defaults to 512) —
Number of output channels used in the ASPP layer for semantic segmentation. List[int], optional, defaults to [6, 12, 18]) —
Dilation (atrous) factors used in the ASPP layer for semantic segmentation. float, optional, defaults to 0.1) —
The dropout ratio for the ASPP layer for semantic segmentation. int, optional, defaults to 255) —
The index that is ignored by the loss function of the semantic segmentation model. List[int], optional, defaults to [2, 4, 3]) —
The number of attention blocks in each MobileViTV2Layer List[int], optional, defaults to [128, 192, 256]) —
The base multiplier for dimensions of attention blocks in each MobileViTV2Layer float, optional, defaults to 1.0) —
The width multiplier for MobileViTV2. int, optional, defaults to 2) —
The FFN multiplier for MobileViTV2. float, optional, defaults to 0.0) —
The dropout in the attention layer. float, optional, defaults to 0.0) —
The dropout between FFN layers. This is the configuration class to store the configuration of a MobileViTV2Model. It is used to instantiate a MobileViTV2 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 MobileViTV2 apple/mobilevitv2-1.0 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 MobileViTV2Config, MobileViTV2Model
>>> # Initializing a mobilevitv2-small style configuration
>>> configuration = MobileViTV2Config()
>>> # Initializing a model from the mobilevitv2-small style configuration
>>> model = MobileViTV2Model(configuration)
>>> # Accessing the model configuration
>>> configuration = model.config( config: MobileViTV2Config expand_output: bool = True )
Parameters
The bare MobileViTV2 model outputting raw hidden-states without any specific head on top. This model is 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.
( pixel_values: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.modeling_outputs.BaseModelOutputWithPoolingAndNoAttention or tuple(torch.FloatTensor)
Parameters
torch.FloatTensor of shape (batch_size, num_channels, height, width)) —
Pixel values. Pixel values can be obtained using AutoImageProcessor. See
MobileViTImageProcessor.call() for details. bool, optional) —
Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for
more detail. bool, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. Returns
transformers.modeling_outputs.BaseModelOutputWithPoolingAndNoAttention or tuple(torch.FloatTensor)
A transformers.modeling_outputs.BaseModelOutputWithPoolingAndNoAttention 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 (MobileViTV2Config) and inputs.
last_hidden_state (torch.FloatTensor of shape (batch_size, num_channels, height, width)) — 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 after a pooling operation on the spatial dimensions.
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, num_channels, height, width).
Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
The MobileViTV2Model 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:
>>> from transformers import AutoImageProcessor, MobileViTV2Model
>>> import torch
>>> from datasets import load_dataset
>>> dataset = load_dataset("huggingface/cats-image")
>>> image = dataset["test"]["image"][0]
>>> image_processor = AutoImageProcessor.from_pretrained("apple/mobilevitv2-1.0-imagenet1k-256")
>>> model = MobileViTV2Model.from_pretrained("apple/mobilevitv2-1.0-imagenet1k-256")
>>> inputs = image_processor(image, return_tensors="pt")
>>> with torch.no_grad():
... outputs = model(**inputs)
>>> last_hidden_states = outputs.last_hidden_state
>>> list(last_hidden_states.shape)
[1, 512, 8, 8]( config: MobileViTV2Config )
Parameters
MobileViTV2 model with an image classification head on top (a linear layer on top of the pooled features), e.g. for ImageNet.
This model is 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.
( pixel_values: Optional = None output_hidden_states: Optional = None labels: Optional = None return_dict: Optional = None ) → transformers.modeling_outputs.ImageClassifierOutputWithNoAttention or tuple(torch.FloatTensor)
Parameters
torch.FloatTensor of shape (batch_size, num_channels, height, width)) —
Pixel values. Pixel values can be obtained using AutoImageProcessor. See
MobileViTImageProcessor.call() for details. bool, optional) —
Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for
more detail. bool, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. 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). Returns
transformers.modeling_outputs.ImageClassifierOutputWithNoAttention or tuple(torch.FloatTensor)
A transformers.modeling_outputs.ImageClassifierOutputWithNoAttention 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 (MobileViTV2Config) and inputs.
torch.FloatTensor of shape (1,), optional, returned when labels is provided) — Classification (or regression if config.num_labels==1) loss.torch.FloatTensor of shape (batch_size, config.num_labels)) — Classification (or regression if config.num_labels==1) scores (before SoftMax).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, num_channels, height, width). Hidden-states (also
called feature maps) of the model at the output of each stage.The MobileViTV2ForImageClassification 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:
>>> from transformers import AutoImageProcessor, MobileViTV2ForImageClassification
>>> import torch
>>> from datasets import load_dataset
>>> dataset = load_dataset("huggingface/cats-image")
>>> image = dataset["test"]["image"][0]
>>> image_processor = AutoImageProcessor.from_pretrained("apple/mobilevitv2-1.0-imagenet1k-256")
>>> model = MobileViTV2ForImageClassification.from_pretrained("apple/mobilevitv2-1.0-imagenet1k-256")
>>> 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])
tabby, tabby cat( config: MobileViTV2Config )
Parameters
MobileViTV2 model with a semantic segmentation head on top, e.g. for Pascal VOC.
This model is 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.
( pixel_values: Optional = None labels: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.modeling_outputs.SemanticSegmenterOutput or tuple(torch.FloatTensor)
Parameters
torch.FloatTensor of shape (batch_size, num_channels, height, width)) —
Pixel values. Pixel values can be obtained using AutoImageProcessor. See
MobileViTImageProcessor.call() for details. bool, optional) —
Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for
more detail. bool, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. torch.LongTensor of shape (batch_size, height, width), optional) —
Ground truth semantic segmentation maps for computing the loss. Indices should be in [0, ..., config.num_labels - 1]. If config.num_labels > 1, a classification loss is computed (Cross-Entropy). Returns
transformers.modeling_outputs.SemanticSegmenterOutput or tuple(torch.FloatTensor)
A transformers.modeling_outputs.SemanticSegmenterOutput 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 (MobileViTV2Config) 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, logits_height, logits_width)) — Classification scores for each pixel.
The logits returned do not necessarily have the same size as the pixel_values passed as inputs. This is
to avoid doing two interpolations and lose some quality when a user needs to resize the logits to the
original image size as post-processing. You should always check your logits shape and resize as needed.
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, patch_size, 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, patch_size, sequence_length).
Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
The MobileViTV2ForSemanticSegmentation 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
>>> import torch
>>> from PIL import Image
>>> from transformers import AutoImageProcessor, MobileViTV2ForSemanticSegmentation
>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)
>>> image_processor = AutoImageProcessor.from_pretrained("apple/mobilevitv2-1.0-imagenet1k-256")
>>> model = MobileViTV2ForSemanticSegmentation.from_pretrained("apple/mobilevitv2-1.0-imagenet1k-256")
>>> inputs = image_processor(images=image, return_tensors="pt")
>>> with torch.no_grad():
... outputs = model(**inputs)
>>> # logits are of shape (batch_size, num_labels, height, width)
>>> logits = outputs.logits