Swin Transformer V2
개요
Swin Transformer V2는 Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo가 제안한 논문 Swin Transformer V2: Scaling Up Capacity and Resolution에서 소개되었습니다.
논문의 초록은 다음과 같습니다:
대규모 NLP 모델들은 언어 작업에서의 성능을 크게 향상하며, 성능이 포화하는 징후를 보이지 않습니다. 또한, 사람과 유사한 few-shot 학습 능력을 보여줍니다. 이 논문은 대규모 모델을 컴퓨터 비전 분야에서 탐구하고자 합니다. 대형 비전 모델을 훈련하고 적용하는 데 있어 세 가지 주요 문제를 다룹니다: 훈련 불안정성, 사전 학습과 파인튜닝 간의 해상도 차이, 그리고 레이블이 달린 데이터에 대한 높은 요구입니다. 세 가지 주요 기법을 제안합니다: 1) 훈련 안정성을 개선하기 위한 residual-post-norm 방법과 cosine attention의 결합; 2) 저해상도 이미지로 사전 학습된 모델을 고해상도 입력으로 전이할 수 있는 log-spaced continuous position bias 방법; 3) 레이블이 달린 방대한 이미지의 필요성을 줄이기 위한 self-supervised 사전 학습 방법인 SimMIM입니다. 이러한 기법들을 통해 30억 개의 파라미터를 가진 Swin Transformer V2 모델을 성공적으로 훈련하였으며, 이는 현재까지 가장 크고 고밀도의 비전 모델로, 최대 1,536×1,536 해상도의 이미지를 다룰 수 있습니다. 이 모델은 ImageNet-V2 이미지 분류, COCO 객체 탐지, ADE20K 의미론적 분할, Kinetics-400 비디오 행동 분류 등 네 가지 대표적인 비전 작업에서 새로운 성능 기록을 세웠습니다. 또한, 우리의 훈련은 Google의 billion-level 비전 모델과 비교해 40배 적은 레이블이 달린 데이터와 40배 적은 훈련 시간으로 이루어졌다는 점에서 훨씬 더 효율적입니다.
이 모델은 nandwalritik이 기여하였습니다. 원본 코드는 여기에서 확인할 수 있습니다.
리소스
Swin Transformer v2의 사용을 도울 수 있는 Hugging Face 및 커뮤니티(🌎로 표시)의 공식 자료 목록입니다.
- Swinv2ForImageClassification은 이 예제 스크립트와 노트북을 통해 지원됩니다.
- 관련 자료: 이미지 분류 작업 가이드
또한:
- Swinv2ForMaskedImageModeling는 이 예제 스크립트를 통해 지원됩니다.
새로운 자료를 추가하고 싶으시다면, 언제든지 Pull Request를 열어주세요! 저희가 검토해 드릴게요. 이때, 추가하는 자료는 기존 자료와 중복되지 않고 새로운 내용을 보여주는 자료여야 합니다.
Swinv2Config
class transformers.Swinv2Config
< source >( image_size = 224 patch_size = 4 num_channels = 3 embed_dim = 96 depths = [2, 2, 6, 2] num_heads = [3, 6, 12, 24] window_size = 7 pretrained_window_sizes = [0, 0, 0, 0] mlp_ratio = 4.0 qkv_bias = True hidden_dropout_prob = 0.0 attention_probs_dropout_prob = 0.0 drop_path_rate = 0.1 hidden_act = 'gelu' use_absolute_embeddings = False initializer_range = 0.02 layer_norm_eps = 1e-05 encoder_stride = 32 out_features = None out_indices = None **kwargs )
Parameters
- image_size (
int, optional, defaults to 224) — The size (resolution) of each image. - patch_size (
int, optional, defaults to 4) — The size (resolution) of each patch. - num_channels (
int, optional, defaults to 3) — The number of input channels. - embed_dim (
int, optional, defaults to 96) — Dimensionality of patch embedding. - depths (
list(int), optional, defaults to[2, 2, 6, 2]) — Depth of each layer in the Transformer encoder. - num_heads (
list(int), optional, defaults to[3, 6, 12, 24]) — Number of attention heads in each layer of the Transformer encoder. - window_size (
int, optional, defaults to 7) — Size of windows. - pretrained_window_sizes (
list(int), optional, defaults to[0, 0, 0, 0]) — Size of windows during pretraining. - mlp_ratio (
float, optional, defaults to 4.0) — Ratio of MLP hidden dimensionality to embedding dimensionality. - qkv_bias (
bool, optional, defaults toTrue) — Whether or not a learnable bias should be added to the queries, keys and values. - hidden_dropout_prob (
float, optional, defaults to 0.0) — The dropout probability for all fully connected layers in the embeddings and encoder. - attention_probs_dropout_prob (
float, optional, defaults to 0.0) — The dropout ratio for the attention probabilities. - drop_path_rate (
float, optional, defaults to 0.1) — Stochastic depth rate. - hidden_act (
strorfunction, optional, defaults to"gelu") — The non-linear activation function (function or string) in the encoder. If string,"gelu","relu","selu"and"gelu_new"are supported. - use_absolute_embeddings (
bool, optional, defaults toFalse) — Whether or not to add absolute position embeddings to the patch embeddings. - 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-05) — The epsilon used by the layer normalization layers. - encoder_stride (
int, optional, defaults to 32) — Factor to increase the spatial resolution by in the decoder head for masked image modeling. - out_features (
List[str], optional) — If used as backbone, list of features to output. Can be any of"stem","stage1","stage2", etc. (depending on how many stages the model has). If unset andout_indicesis set, will default to the corresponding stages. If unset andout_indicesis unset, will default to the last stage. - out_indices (
List[int], optional) — If used as backbone, list of indices of features to output. Can be any of 0, 1, 2, etc. (depending on how many stages the model has). If unset andout_featuresis set, will default to the corresponding stages. If unset andout_featuresis unset, will default to the last stage.
This is the configuration class to store the configuration of a Swinv2Model. It is used to instantiate a Swin Transformer v2 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 Swin Transformer v2 microsoft/swinv2-tiny-patch4-window8-256 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 Swinv2Config, Swinv2Model
>>> # Initializing a Swinv2 microsoft/swinv2-tiny-patch4-window8-256 style configuration
>>> configuration = Swinv2Config()
>>> # Initializing a model (with random weights) from the microsoft/swinv2-tiny-patch4-window8-256 style configuration
>>> model = Swinv2Model(configuration)
>>> # Accessing the model configuration
>>> configuration = model.configSwinv2Model
class transformers.Swinv2Model
< source >( config add_pooling_layer = True use_mask_token = False )
Parameters
- config (Swinv2Config) — 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.
The bare Swinv2 Model transformer outputting raw hidden-states without any specific head on top. This model is a PyTorch torch.nn.Module sub-class. 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: Optional = None bool_masked_pos: Optional = None head_mask: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None interpolate_pos_encoding: bool = False return_dict: Optional = None ) → transformers.models.swinv2.modeling_swinv2.Swinv2ModelOutput or tuple(torch.FloatTensor)
Parameters
- pixel_values (
torch.FloatTensorof shape(batch_size, num_channels, height, width)) — Pixel values. Pixel values can be obtained using AutoImageProcessor. See ViTImageProcessor.call() for details. - head_mask (
torch.FloatTensorof 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. Seeattentionsunder returned tensors for more detail. - output_hidden_states (
bool, optional) — Whether or not to return the hidden states of all layers. Seehidden_statesunder returned tensors for more detail. - interpolate_pos_encoding (
bool, optional, defaultFalse) — Whether to interpolate the pre-trained position encodings. - return_dict (
bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple. - bool_masked_pos (
torch.BoolTensorof shape(batch_size, num_patches), optional) — Boolean masked positions. Indicates which patches are masked (1) and which aren’t (0).
Returns
transformers.models.swinv2.modeling_swinv2.Swinv2ModelOutput or tuple(torch.FloatTensor)
A transformers.models.swinv2.modeling_swinv2.Swinv2ModelOutput 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 (Swinv2Config) and inputs.
-
last_hidden_state (
torch.FloatTensorof shape(batch_size, sequence_length, hidden_size)) — Sequence of hidden-states at the output of the last layer of the model. -
pooler_output (
torch.FloatTensorof shape(batch_size, hidden_size), optional, returned whenadd_pooling_layer=Trueis passed) — Average pooling of the last layer hidden-state. -
hidden_states (
tuple(torch.FloatTensor), optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each stage) of shape(batch_size, sequence_length, hidden_size).Hidden-states of the model at the output of each layer plus the initial embedding outputs.
-
attentions (
tuple(torch.FloatTensor), optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.FloatTensor(one for each stage) 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.
-
reshaped_hidden_states (
tuple(torch.FloatTensor), optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each stage) of shape(batch_size, hidden_size, height, width).Hidden-states of the model at the output of each layer plus the initial embedding outputs reshaped to include the spatial dimensions.
The Swinv2Model 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, Swinv2Model
>>> import torch
>>> from datasets import load_dataset
>>> dataset = load_dataset("huggingface/cats-image", trust_remote_code=True)
>>> image = dataset["test"]["image"][0]
>>> image_processor = AutoImageProcessor.from_pretrained("microsoft/swinv2-tiny-patch4-window8-256")
>>> model = Swinv2Model.from_pretrained("microsoft/swinv2-tiny-patch4-window8-256")
>>> 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, 64, 768]Swinv2ForMaskedImageModeling
class transformers.Swinv2ForMaskedImageModeling
< source >( config )
Parameters
- config (Swinv2Config) — 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.
Swinv2 Model with a decoder on top for masked image modeling, as proposed in SimMIM.
Note that we provide a script to pre-train this model on custom data in our examples directory.
This model is a PyTorch torch.nn.Module sub-class. 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: Optional = None bool_masked_pos: Optional = None head_mask: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None interpolate_pos_encoding: bool = False return_dict: Optional = None ) → transformers.models.swinv2.modeling_swinv2.Swinv2MaskedImageModelingOutput or tuple(torch.FloatTensor)
Parameters
- pixel_values (
torch.FloatTensorof shape(batch_size, num_channels, height, width)) — Pixel values. Pixel values can be obtained using AutoImageProcessor. See ViTImageProcessor.call() for details. - head_mask (
torch.FloatTensorof 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. Seeattentionsunder returned tensors for more detail. - output_hidden_states (
bool, optional) — Whether or not to return the hidden states of all layers. Seehidden_statesunder returned tensors for more detail. - interpolate_pos_encoding (
bool, optional, defaultFalse) — Whether to interpolate the pre-trained position encodings. - return_dict (
bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple. - bool_masked_pos (
torch.BoolTensorof shape(batch_size, num_patches)) — Boolean masked positions. Indicates which patches are masked (1) and which aren’t (0).
Returns
transformers.models.swinv2.modeling_swinv2.Swinv2MaskedImageModelingOutput or tuple(torch.FloatTensor)
A transformers.models.swinv2.modeling_swinv2.Swinv2MaskedImageModelingOutput 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 (Swinv2Config) and inputs.
-
loss (
torch.FloatTensorof shape(1,), optional, returned whenbool_masked_posis provided) — Masked image modeling (MLM) loss. -
reconstruction (
torch.FloatTensorof shape(batch_size, num_channels, height, width)) — Reconstructed pixel values. -
hidden_states (
tuple(torch.FloatTensor), optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each stage) of shape(batch_size, sequence_length, hidden_size).Hidden-states of the model at the output of each layer plus the initial embedding outputs.
-
attentions (
tuple(torch.FloatTensor), optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.FloatTensor(one for each stage) 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.
-
reshaped_hidden_states (
tuple(torch.FloatTensor), optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each stage) of shape(batch_size, hidden_size, height, width).Hidden-states of the model at the output of each layer plus the initial embedding outputs reshaped to include the spatial dimensions.
The Swinv2ForMaskedImageModeling 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 AutoImageProcessor, Swinv2ForMaskedImageModeling
>>> import torch
>>> from PIL import Image
>>> import requests
>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)
>>> image_processor = AutoImageProcessor.from_pretrained("microsoft/swinv2-tiny-patch4-window8-256")
>>> model = Swinv2ForMaskedImageModeling.from_pretrained("microsoft/swinv2-tiny-patch4-window8-256")
>>> num_patches = (model.config.image_size // model.config.patch_size) ** 2
>>> pixel_values = image_processor(images=image, return_tensors="pt").pixel_values
>>> # create random boolean mask of shape (batch_size, num_patches)
>>> bool_masked_pos = torch.randint(low=0, high=2, size=(1, num_patches)).bool()
>>> outputs = model(pixel_values, bool_masked_pos=bool_masked_pos)
>>> loss, reconstructed_pixel_values = outputs.loss, outputs.reconstruction
>>> list(reconstructed_pixel_values.shape)
[1, 3, 256, 256]Swinv2ForImageClassification
class transformers.Swinv2ForImageClassification
< source >( config )
Parameters
- config (Swinv2Config) — 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.
Swinv2 Model transformer with an image classification head on top (a linear layer on top of the final hidden state of the [CLS] token) e.g. for ImageNet.
Note that it’s possible to fine-tune SwinV2 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 a PyTorch torch.nn.Module sub-class. 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: Optional = None head_mask: Optional = None labels: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None interpolate_pos_encoding: bool = False return_dict: Optional = None ) → transformers.models.swinv2.modeling_swinv2.Swinv2ImageClassifierOutput or tuple(torch.FloatTensor)
Parameters
- pixel_values (
torch.FloatTensorof shape(batch_size, num_channels, height, width)) — Pixel values. Pixel values can be obtained using AutoImageProcessor. See ViTImageProcessor.call() for details. - head_mask (
torch.FloatTensorof 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. Seeattentionsunder returned tensors for more detail. - output_hidden_states (
bool, optional) — Whether or not to return the hidden states of all layers. Seehidden_statesunder returned tensors for more detail. - interpolate_pos_encoding (
bool, optional, defaultFalse) — Whether to interpolate the pre-trained position encodings. - return_dict (
bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple. - labels (
torch.LongTensorof shape(batch_size,), optional) — Labels for computing the image classification/regression loss. Indices should be in[0, ..., config.num_labels - 1]. Ifconfig.num_labels == 1a regression loss is computed (Mean-Square loss), Ifconfig.num_labels > 1a classification loss is computed (Cross-Entropy).
Returns
transformers.models.swinv2.modeling_swinv2.Swinv2ImageClassifierOutput or tuple(torch.FloatTensor)
A transformers.models.swinv2.modeling_swinv2.Swinv2ImageClassifierOutput 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 (Swinv2Config) and inputs.
-
loss (
torch.FloatTensorof shape(1,), optional, returned whenlabelsis provided) — Classification (or regression if config.num_labels==1) loss. -
logits (
torch.FloatTensorof shape(batch_size, config.num_labels)) — Classification (or regression if config.num_labels==1) scores (before SoftMax). -
hidden_states (
tuple(torch.FloatTensor), optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each stage) of shape(batch_size, sequence_length, hidden_size).Hidden-states of the model at the output of each layer plus the initial embedding outputs.
-
attentions (
tuple(torch.FloatTensor), optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.FloatTensor(one for each stage) 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.
-
reshaped_hidden_states (
tuple(torch.FloatTensor), optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each stage) of shape(batch_size, hidden_size, height, width).Hidden-states of the model at the output of each layer plus the initial embedding outputs reshaped to include the spatial dimensions.
The Swinv2ForImageClassification 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, Swinv2ForImageClassification
>>> import torch
>>> from datasets import load_dataset
>>> dataset = load_dataset("huggingface/cats-image", trust_remote_code=True)
>>> image = dataset["test"]["image"][0]
>>> image_processor = AutoImageProcessor.from_pretrained("microsoft/swinv2-tiny-patch4-window8-256")
>>> model = Swinv2ForImageClassification.from_pretrained("microsoft/swinv2-tiny-patch4-window8-256")
>>> 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])
Egyptian cat