ViT-Calibrator: Decision Stream Calibration for Vision Transformer
- URL: http://arxiv.org/abs/2304.04354v2
- Date: Fri, 5 May 2023 13:41:40 GMT
- Title: ViT-Calibrator: Decision Stream Calibration for Vision Transformer
- Authors: Lin Chen, Zhijie Jia, Tian Qiu, Lechao Cheng, Jie Lei, Zunlei Feng,
Mingli Song
- Abstract summary: We propose a new paradigm dubbed Decision Stream that boosts the performance of general Vision Transformers.
We shed light on the information propagation mechanism in the learning procedure by exploring the correlation between different tokens and the relevance coefficient of multiple dimensions.
- Score: 49.60474757318486
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: A surge of interest has emerged in utilizing Transformers in diverse vision
tasks owing to its formidable performance. However, existing approaches
primarily focus on optimizing internal model architecture designs that often
entail significant trial and error with high burdens. In this work, we propose
a new paradigm dubbed Decision Stream Calibration that boosts the performance
of general Vision Transformers. To achieve this, we shed light on the
information propagation mechanism in the learning procedure by exploring the
correlation between different tokens and the relevance coefficient of multiple
dimensions. Upon further analysis, it was discovered that 1) the final decision
is associated with tokens of foreground targets, while token features of
foreground target will be transmitted into the next layer as much as possible,
and the useless token features of background area will be eliminated gradually
in the forward propagation. 2) Each category is solely associated with specific
sparse dimensions in the tokens. Based on the discoveries mentioned above, we
designed a two-stage calibration scheme, namely ViT-Calibrator, including token
propagation calibration stage and dimension propagation calibration stage.
Extensive experiments on commonly used datasets show that the proposed approach
can achieve promising results. The source codes are given in the supplements.
Related papers
- CLAMP-ViT: Contrastive Data-Free Learning for Adaptive Post-Training Quantization of ViTs [6.456189487006878]
We present CLAMP-ViT, a data-free post-training quantization method for vision transformers (ViTs)
We identify the limitations of recent techniques, notably their inability to leverage meaningful inter-patch relationships.
CLAMP-ViT employs a two-stage approach, cyclically adapting between data generation and model quantization.
arXiv Detail & Related papers (2024-07-07T05:39:25Z) - Visual Prompt Tuning in Null Space for Continual Learning [51.96411454304625]
Existing prompt-tuning methods have demonstrated impressive performances in continual learning (CL)
This paper aims to learn each task by tuning the prompts in the direction orthogonal to the subspace spanned by previous tasks' features.
In practice, an effective null-space-based approximation solution has been proposed to implement the prompt gradient projection.
arXiv Detail & Related papers (2024-06-09T05:57:40Z) - LRP-QViT: Mixed-Precision Vision Transformer Quantization via Layer-wise
Relevance Propagation [0.0]
We introduce LRP-QViT, an explainability-based method for assigning mixed-precision bit allocations to different layers based on their importance during classification.
Our experimental findings demonstrate that both our fixed-bit and mixed-bit post-training quantization methods surpass existing models in the context of 4-bit and 6-bit quantization.
arXiv Detail & Related papers (2024-01-20T14:53:19Z) - Denoising Vision Transformers [43.03068202384091]
We propose a two-stage denoising approach, termed Denoising Vision Transformers (DVT)
In the first stage, we separate the clean features from those contaminated by positional artifacts by enforcing cross-view feature consistency with neural fields on a per-image basis.
In the second stage, we train a lightweight transformer block to predict clean features from raw ViT outputs, leveraging the derived estimates of the clean features as supervision.
arXiv Detail & Related papers (2024-01-05T18:59:52Z) - BTranspose: Bottleneck Transformers for Human Pose Estimation with
Self-Supervised Pre-Training [0.304585143845864]
In this paper, we consider the recently proposed Bottleneck Transformers, which combine CNN and multi-head self attention (MHSA) layers effectively.
We consider different backbone architectures and pre-train them using the DINO self-supervised learning method.
Experiments show that our model achieves an AP of 76.4, which is competitive with other methods such as [1] and has fewer network parameters.
arXiv Detail & Related papers (2022-04-21T15:45:05Z) - Visualizing and Understanding Patch Interactions in Vision Transformer [96.70401478061076]
Vision Transformer (ViT) has become a leading tool in various computer vision tasks.
We propose a novel explainable visualization approach to analyze and interpret the crucial attention interactions among patches for vision transformer.
arXiv Detail & Related papers (2022-03-11T13:48:11Z) - Augmented Shortcuts for Vision Transformers [49.70151144700589]
We study the relationship between shortcuts and feature diversity in vision transformer models.
We present an augmented shortcut scheme, which inserts additional paths with learnable parameters in parallel on the original shortcuts.
Experiments conducted on benchmark datasets demonstrate the effectiveness of the proposed method.
arXiv Detail & Related papers (2021-06-30T09:48:30Z) - Visual Saliency Transformer [127.33678448761599]
We develop a novel unified model based on a pure transformer, Visual Saliency Transformer (VST), for both RGB and RGB-D salient object detection (SOD)
It takes image patches as inputs and leverages the transformer to propagate global contexts among image patches.
Experimental results show that our model outperforms existing state-of-the-art results on both RGB and RGB-D SOD benchmark datasets.
arXiv Detail & Related papers (2021-04-25T08:24:06Z) - Transformers Solve the Limited Receptive Field for Monocular Depth
Prediction [82.90445525977904]
We propose TransDepth, an architecture which benefits from both convolutional neural networks and transformers.
This is the first paper which applies transformers into pixel-wise prediction problems involving continuous labels.
arXiv Detail & Related papers (2021-03-22T18:00:13Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.