Multiple-Exit Tuning: Towards Inference-Efficient Adaptation for Vision Transformer

Parameter-efficient transfer learning (PETL) has shown great potential in adapting a vision transformer (ViT) pre-trained on large-scale datasets to various downstream tasks. Existing studies primarily focus on minimizing the number of learnable parameters. Although these methods are storage-efficient, they allocate excessive computational resources to easy samples, leading to inefficient inference. To address this issue, we introduce an inference-efficient tuning method termed multiple-exit tuning (MET). MET integrates multiple exits into the pre-trained ViT backbone. Since the predictions in ViT are made by a linear classifier, each exit is equipped with a linear prediction head. In inference stage, easy samples will exit at early exits and only hard enough samples will flow to the last exit, thus saving the computational cost for easy samples. MET consists of exit-specific adapters (E-adapters) and graph regularization. E-adapters are designed to extract suitable representations for different exits. To ensure parameter efficiency, all E-adapters share the same down-projection and up-projection matrices. As the performances of linear classifiers are influenced by the relationship among samples, we employ graph regularization to improve the representations fed into the classifiers at early exits. Finally, we conduct extensive experiments to verify the performance of MET. Experimental results show that MET has an obvious advantage over the state-of-the-art methods in terms of both accuracy and inference efficiency.

PanoFree: Tuning-Free Holistic Multi-view Image Generation with Cross-view Self-Guidance

Immersive scene generation, notably panorama creation, benefits significantly from the adaptation of large pre-trained text-to-image (T2I) models for multi-view image generation. Due to the high cost of acquiring multi-view images, tuning-free generation is preferred. However, existing methods are either limited to simple correspondences or require extensive fine-tuning to capture complex ones. We present PanoFree, a novel method for tuning-free multi-view image generation that supports an extensive array of correspondences. PanoFree sequentially generates multi-view images using iterative warping and inpainting, addressing the key issues of inconsistency and artifacts from error accumulation without the need for fine-tuning. It improves error accumulation by enhancing cross-view awareness and refines the warping and inpainting processes via cross-view guidance, risky area estimation and erasing, and symmetric bidirectional guided generation for loop closure, alongside guidance-based semantic and density control for scene structure preservation. In experiments on Planar, 360{\deg}, and Full Spherical Panoramas, PanoFree demonstrates significant error reduction, improves global consistency, and boosts image quality without extra fine-tuning. Compared to existing methods, PanoFree is up to 5x more efficient in time and 3x more efficient in GPU memory usage, and maintains superior diversity of results (2x better in our user study). PanoFree offers a viable alternative to costly fine-tuning or the use of additional pre-trained models. Project website at https://panofree.github.io/.

UniHuman: A Unified Model for Editing Human Images in the Wild

Human image editing includes tasks like changing a person's pose, their clothing, or editing the image according to a text prompt. However, prior work often tackles these tasks separately, overlooking the benefit of mutual reinforcement from learning them jointly. In this paper, we propose UniHuman, a unified model that addresses multiple facets of human image editing in real-world settings. To enhance the model's generation quality and generalization capacity, we leverage guidance from human visual encoders and introduce a lightweight pose-warping module that can exploit different pose representations, accommodating unseen textures and patterns. Furthermore, to bridge the disparity between existing human editing benchmarks with real-world data, we curated 400K high-quality human image-text pairs for training and collected 2K human images for out-of-domain testing, both encompassing diverse clothing styles, backgrounds, and age groups. Experiments on both in-domain and out-of-domain test sets demonstrate that UniHuman outperforms task-specific models by a significant margin. In user studies, UniHuman is preferred by the users in an average of 77% of cases.

Fine-grained Text and Image Guided Point Cloud Completion with CLIP Model

This paper focuses on the recently popular task of point cloud completion guided by multimodal information. Although existing methods have achieved excellent performance by fusing auxiliary images, there are still some deficiencies, including the poor generalization ability of the model and insufficient fine-grained semantic information for extracted features. In this work, we propose a novel multimodal fusion network for point cloud completion, which can simultaneously fuse visual and textual information to predict the semantic and geometric characteristics of incomplete shapes effectively. Specifically, to overcome the lack of prior information caused by the small-scale dataset, we employ a pre-trained vision-language model that is trained with a large amount of image-text pairs. Therefore, the textual and visual encoders of this large-scale model have stronger generalization ability. Then, we propose a multi-stage feature fusion strategy to fuse the textual and visual features into the backbone network progressively. Meanwhile, to further explore the effectiveness of fine-grained text descriptions for point cloud completion, we also build a text corpus with fine-grained descriptions, which can provide richer geometric details for 3D shapes. The rich text descriptions can be used for training and evaluating our network. Extensive quantitative and qualitative experiments demonstrate the superior performance of our method compared to state-of-the-art point cloud completion networks.

Leaf Cultivar Identification via Prototype-enhanced Learning

Plant leaf identification is crucial for biodiversity protection and conservation and has gradually attracted the attention of academia in recent years. Due to the high similarity among different varieties, leaf cultivar recognition is also considered to be an ultra-fine-grained visual classification (UFGVC) task, which is facing a huge challenge. In practice, an instance may be related to multiple varieties to varying degrees, especially in the UFGVC datasets. However, deep learning methods trained on one-hot labels fail to reflect patterns shared across categories and thus perform poorly on this task. To address this issue, we generate soft targets integrated with inter-class similarity information. Specifically, we continuously update the prototypical features for each category and then capture the similarity scores between instances and prototypes accordingly. Original one-hot labels and the similarity scores are incorporated to yield enhanced labels. Prototype-enhanced soft labels not only contain original one-hot label information, but also introduce rich inter-category semantic association information, thus providing more effective supervision for deep model training. Extensive experimental results on public datasets show that our method can significantly improve the performance on the UFGVC task of leaf cultivar identification.

Collecting The Puzzle Pieces: Disentangled Self-Driven Human Pose Transfer by Permuting Textures

Human pose transfer aims to synthesize a new view of a person under a given pose. Recent works achieve this via self-reconstruction, which disentangles pose and texture features from the person image, then combines the two features to reconstruct the person. Such feature-level disentanglement is a difficult and ill-defined problem that could lead to loss of details and unwanted artifacts. In this paper, we propose a self-driven human pose transfer method that permutes the textures at random, then reconstructs the image with a dual branch attention to achieve image-level disentanglement and detail-preserving texture transfer. We find that compared with feature-level disentanglement, image-level disentanglement is more controllable and reliable. Furthermore, we introduce a dual kernel encoder that gives different sizes of receptive fields in order to reduce the noise caused by permutation and thus recover clothing details while aligning pose and textures. Extensive experiments on DeepFashion and Market-1501 shows that our model improves the quality of generated images in terms of FID, LPIPS and SSIM over other self-driven methods, and even outperforming some fully-supervised methods. A user study also shows that among self-driven approaches, images generated by our method are preferred in 72% of cases over prior work.

Supervised Attribute Information Removal and Reconstruction for Image Manipulation

The goal of attribute manipulation is to control specified attribute(s) in given images. Prior work approaches this problem by learning disentangled representations for each attribute that enables it to manipulate the encoded source attributes to the target attributes. However, encoded attributes are often correlated with relevant image content. Thus, the source attribute information can often be hidden in the disentangled features, leading to unwanted image editing effects. In this paper, we propose an Attribute Information Removal and Reconstruction (AIRR) network that prevents such information hiding by learning how to remove the attribute information entirely, creating attribute excluded features, and then learns to directly inject the desired attributes in a reconstructed image. We evaluate our approach on four diverse datasets with a variety of attributes including DeepFashion Synthesis, DeepFashion Fine-grained Attribute, CelebA and CelebA-HQ, where our model improves attribute manipulation accuracy and top-k retrieval rate by 10% on average over prior work. A user study also reports that AIRR manipulated images are preferred over prior work in up to 76% of cases.

A Unified Weight Initialization Paradigm for Tensorial Convolutional Neural Networks

Tensorial Convolutional Neural Networks (TCNNs) have attracted much research attention for their power in reducing model parameters or enhancing the generalization ability. However, exploration of TCNNs is hindered even from weight initialization methods. To be specific, general initialization methods, such as Xavier or Kaiming initialization, usually fail to generate appropriate weights for TCNNs. Meanwhile, although there are ad-hoc approaches for specific architectures (e.g., Tensor Ring Nets), they are not applicable to TCNNs with other tensor decomposition methods (e.g., CP or Tucker decomposition). To address this problem, we propose a universal weight initialization paradigm, which generalizes Xavier and Kaiming methods and can be widely applicable to arbitrary TCNNs. Specifically, we first present the Reproducing Transformation to convert the backward process in TCNNs to an equivalent convolution process. Then, based on the convolution operators in the forward and backward processes, we build a unified paradigm to control the variance of features and gradients in TCNNs. Thus, we can derive fan-in and fan-out initialization for various TCNNs. We demonstrate that our paradigm can stabilize the training of TCNNs, leading to faster convergence and better results.

BViT: Broad Attention based Vision Transformer

Recent works have demonstrated that transformer can achieve promising performance in computer vision, by exploiting the relationship among image patches with self-attention. While they only consider the attention in a single feature layer, but ignore the complementarity of attention in different levels. In this paper, we propose the broad attention to improve the performance by incorporating the attention relationship of different layers for vision transformer, which is called BViT. The broad attention is implemented by broad connection and parameter-free attention. Broad connection of each transformer layer promotes the transmission and integration of information for BViT. Without introducing additional trainable parameters, parameter-free attention jointly focuses on the already available attention information in different layers for extracting useful information and building their relationship. Experiments on image classification tasks demonstrate that BViT delivers state-of-the-art accuracy of 74.8\%/81.6\% top-1 accuracy on ImageNet with 5M/22M parameters. Moreover, we transfer BViT to downstream object recognition benchmarks to achieve 98.9\% and 89.9\% on CIFAR10 and CIFAR100 respectively that exceed ViT with fewer parameters. For the generalization test, the broad attention in Swin Transformer and T2T-ViT also bring an improvement of more than 1\%. To sum up, broad attention is promising to promote the performance of attention based models. Code and pre-trained models are available at https://github.com/DRL-CASIA/Broad_ViT.

Stacked BNAS: Rethinking Broad Convolutional Neural Network for Neural Architecture Search

Different from other deep scalable architecture based NAS approaches, Broad Neural Architecture Search (BNAS) proposes a broad one which consists of convolution and enhancement blocks, dubbed Broad Convolutional Neural Network (BCNN) as search space for amazing efficiency improvement. BCNN reuses the topologies of cells in convolution block, so that BNAS can employ few cells for efficient search. Moreover, multi-scale feature fusion and knowledge embedding are proposed to improve the performance of BCNN with shallow topology. However, BNAS suffers some drawbacks: 1) insufficient representation diversity for feature fusion and enhancement, and 2) time consuming of knowledge embedding design by human expert. In this paper, we propose Stacked BNAS whose search space is a developed broad scalable architecture named Stacked BCNN, with better performance than BNAS. On the one hand, Stacked BCNN treats mini-BCNN as the basic block to preserve comprehensive representation and deliver powerful feature extraction ability. On the other hand, we propose Knowledge Embedding Search (KES) to learn appropriate knowledge embeddings. Experimental results show that 1) Stacked BNAS obtains better performance than BNAS, 2) KES contributes to reduce the parameters of learned architecture with satisfactory performance, and 3) Stacked BNAS delivers state-of-the-art efficiency of 0.02 GPU days.