Diffusion Transformer (DiT), an emerging diffusion model for visual generation, has demonstrated superior performance but suffers from substantial computational costs. Our investigations reveal that these costs primarily stem from the static inference paradigm, which inevitably introduces redundant computation in certain diffusion timesteps and spatial regions. To overcome this inefficiency, we propose Dynamic Diffusion Transformer (DyDiT), an architecture that dynamically adjusts its computation along both timestep and spatial dimensions. Specifically, we introduce a Timestep-wise Dynamic Width (TDW) approach that adapts model width conditioned on the generation timesteps. In addition, we design a Spatial-wise Dynamic Token (SDT) strategy to avoid redundant computation at unnecessary spatial locations. TDW and SDT can be seamlessly integrated into DiT and significantly accelerate the generation process. Building on these designs, we present an extended version, DyDiT++, with improvements in three key aspects. First, it extends the generation mechanism of DyDiT beyond diffusion to flow matching, demonstrating that our method can also accelerate flow-matching-based generation, enhancing its versatility. Furthermore, we enhance DyDiT to tackle more complex visual generation tasks, including video generation and text-to-image generation, thereby broadening its real-world applications. Finally, to address the high cost of full fine-tuning and democratize technology access, we investigate the feasibility of training DyDiT in a parameter-efficient manner and introduce timestep-based dynamic LoRA (TD-LoRA). Extensive experiments on diverse visual generation models, including DiT, SiT, Latte, and FLUX, demonstrate the effectiveness of DyDiT++. Remarkably, with $< $<3% additional fine-tuning iterations, our approach reduces the FLOPs of DiT-XL by 51%, yielding 1.73× realistic speedup on hardware, and achieves a competitive FID score of 2.07 on ImageNet.

Pre-trained Vision-Language Models (VLMs) are often used to tackle the challenging task of Open-vocabulary Segmentation (OVS). To preserve the valuable pre-trained knowledge of VLM-based mask classifiers, most existing approaches freeze their parameters during training. However, our comprehensive analysis identifies a previously overlooked limitation: the performance of OVS is primarily constrained by mask classification. Specifically, VLMs pre-trained using globally pooled image-text representations often fail to capture localized, region-specific semantics necessary for accurate segmentation. This discovery motivates us to improve the fine-grained alignment between word-level text features and pixel-level image features extracted by VLMs. To this end, we propose the Fine-grained Semantic Reconstruction (FiSeR), a novel auxiliary task designed to enrich the spatial semantic detail of visual features. FiSeR trains the model to predict a randomly masked target class label using the image features and the remaining unmasked text. This encourages the model to link the specific words to the corresponding image regions, improving its ability to recognize and segment objects at the region level. FiSeR is broadly applicable and can be incorporated into various VLM-based segmentation models to improve their performance. Additionally, we introduce the Text-guided Visual Aligner (TeVA), a lightweight network module that injects relevant fine-grained semantics from the text information early in the visual encoding process. This enables the model to condition its visual processing on the target text categories from the beginning, improving its ability to associate text with the correct spatial regions. Collectively, these innovations culminate in our proposed framework FOV-Seg. Notably, FOV-Seg achieves new state-of-the-art results across multiple representative OVS benchmarks, improving performance consistently and reducing training costs by nearly 5× compared to previous best methods. Our code and data will be released.

Vision-language models (VLMs) have shown remarkable success across various multi-modal tasks, yet large VLMs encounter significant efficiency challenges due to processing numerous visual tokens. A promising approach to accelerating large VLM inference is using partial information, such as attention maps from specific layers, to assess token importance and prune less essential tokens. However, our study reveals three key insights: (i) Partial attention information is insufficient for accurately identifying critical visual tokens, resulting in suboptimal performance, especially at low token retention ratios; (ii) Global attention information, such as the attention map aggregated across all layers, more effectively preserves essential tokens and maintains comparable performance under aggressive pruning. However, the attention maps from all layers require a full inference pass, which increases computational load and is therefore impractical in existing methods; and (iii) The global attention map aggregated from a small VLM closely resembles that of a large VLM, suggesting an efficient alternative. Based on these findings, we introduce a training-free method, Small VLM Guidance for accelerating Large VLMs (SGL). Specifically, we employ the attention map aggregated from a small VLM to guide visual token pruning in a large VLM. Additionally, an early exiting mechanism is developed to fully use the small VLM’s predictions, dynamically invoking the larger VLM only when necessary, yielding a superior trade-off between accuracy and computation. Extensive evaluations across 11 benchmarks demonstrate the effectiveness and generalizability of SGL, achieving up to 91% pruning ratio for visual tokens while retaining competitive performance. The code is publicly available at https://github.com/NUS-HPC-AI-Lab/SGL.

Diffusion Transformer (DiT), an emerging diffusion model for image generation, has demonstrated superior performance but suffers from substantial computational costs. Our investigations reveal that these costs stem from the static inference paradigm, which inevitably introduces redundant computation in certain diffusion timesteps and spatial regions. To address this inefficiency, we propose Dynamic Diffusion Transformer (DyDiT), an architecture that dynamically adjusts its compu- tation along both timestep and spatial dimensions during generation. Specifically, we introduce a Timestep-wise Dynamic Width (TDW) approach that adapts model width conditioned on the generation timesteps. In addition, we design a Spatial- wise Dynamic Token (SDT) strategy to avoid redundant computation at unnecessary spatial locations. Extensive experiments on various datasets and different-sized models verify the superiority of DyDiT. Notably, with <3% additional fine-tuning it- erations, our method reduces the FLOPs of DiT-XL by 51%, accelerates generation by 1.73×, and achieves a competitive FID score of 2.07 on ImageNet.

Diffusion generative models have become the standard for producing high-quality, coherent video content, yet their slow inference speeds and high computational demands hinder practical deployment. Although both quantization and sparsity can independently accelerate inference while maintaining generation quality, naively combining these techniques in existing training-free approaches leads to significant performance degradation, as they fail to achieve proper joint optimization. We introduce FPSAttention, a novel training-aware co-design of FP8 quantization and Sparsity for video generation, with a focus on the 3D bi-directional attention mechanism. Our approach features three key innovations: 1) A unified 3D tile-wise granularity that simultaneously supports both quantization and sparsity. 2) A denoising step-aware strategy that adapts to the noise schedule, addressing the strong correlation between quantization/sparsity errors and denoising steps. 3) A native, hardware-friendly kernel that leverages FlashAttention and is implemented with optimized Hopper architecture features, enabling highly efficient execution. Trained on Wan2.1's 1.3B and 14B models and evaluated on the vBench benchmark, FPSAttention achieves a 7.09$\times$ kernel speedup for attention operations and a 4.96$\times$ end-to-end speedup for video generation compared to the BF16 baseline at 720p resolution—without sacrificing generation quality.

Referring Image Segmentation (RIS) is a challenging task that requires an algorithm to segment objects referred by free-form language expressions. Despite significant progress in recent years, most state-of-the-art (SOTA) methods still suffer from considerable language-image modality gap at the pixel and word level. These methods generally 1) rely on sentence-level language features for languageimage alignment and 2) lack explicit training supervision for fine-grained visual grounding. Consequently, they exhibit weak object-level correspondence between visual and language features. Without well-grounded features, prior methods struggle to understand complex expressions that require strong reasoning over relationships among multiple objects, especially when dealing with rarely used or ambiguous clauses. To tackle this challenge, we introduce a novel Mask Grounding auxiliary task that significantly improves visual grounding within language features, by explicitly teaching the model to learn fine-grained correspondence between masked textual tokens and their matching visual objects. Mask Grounding can be directly used on prior RIS methods and consistently bring improvements. Furthermore, to holistically address the modality gap, we also design a cross-modal alignment loss and an accompanying alignment module. These additions work synergistically with Mask Grounding. With all these techniques, our comprehensive approach culminates in MagNet (Mask-grounded Network), an architecture that significantly outperforms prior arts on three key benchmarks (Ref CO CO, RefCOCO+ and G-Ref), demonstrating our method's effectiveness in addressing current limitations of RIS algorithms. Our code and pre-trained weights will be released.

Generalized Referring Expression Segmentation (GRES) extends the scope of classic RES to refer to multiple ob-jects in one expression or identify the empty targets absent in the image. GRES poses challenges in modeling the com-plex spatial relationships of the instances in the image and identifying non-existing referents. Multimodal Large Language Models (MLLMs) have recently shown tremendous progress in these complicated vision-language tasks. Con-necting Large Language Models (LLMs) and vision models, MLLMs are proficient in understanding contexts with visual inputs. Among them, LISA, as a representative, adopts a special [SEG] token to prompt a segmentation mask de-coder, e.g., SAM, to enable MLLMs in the RES task. How-ever, existing solutions to GRES remain unsatisfactory since current segmentation MLLMs cannot correctly handle the cases where users might reference multiple subjects in a singular prompt or provide descriptions incongruent with any image target. In this paper, we propose Generalized Segmentation Vision Assistant (GSVA) to address this gap. Specifically, GSVA reuses the [SEG] token to prompt the segmentation model towards supporting multiple mask ref-erences simultaneously and innovatively learns to generate a [REJ] token to reject the null targets explicitly. Ex-periments validate GSVA's efficacy in resolving the GRES issue, marking a notable enhancement and setting a new record on the GRES benchmark gRefCOCO dataset. GSVA also proves effective across various classic referring seg-mentation and comprehension tasks. Code is available at https://github.com/LeapLabTHU/GSVA.

Recent advancements in semi-supervised learning have focused on a more realistic yet challenging task: addressing imbalances in labeled data while the class distribution of unlabeled data remains both unknown and potentially mismatched. Current approaches in this sphere often presuppose rigid assumptions regarding the class distribution of unlabeled data, thereby limiting the adaptability of models to only certain distribution ranges. In this study, we propose a novel approach, introducing a highly adaptable framework, designated as SimPro, which does not rely on any predefined assumptions about the distribution of unlabeled data. Our framework, grounded in a probabilistic model, innovatively refines the expectation-maximization (EM) method by separating the modeling of conditional and marginal class distributions. This separation facilitates a closed-form solution for class distribution estimation during the maximization phase, leading to the formulation of a Bayes classifier. The Bayes classifier, in turn, enhances the quality of pseudo-labels in the expectation phase. Remarkably, the SimPro framework is not only straightforward to implement but also comes with theoretical guarantees. Moreover, we introduce two novel class distributions broadening the scope of the evaluation. Our method showcases consistent state-of-the-art performance across diverse benchmarks and data distribution scenarios. benchmarks and data distribution scenarios. Our code is available at https://github.com/LeapLabTHU/SimPro.

Widely adopted in modern Vision Transformer designs, Softmax attention can effectively capture long-range visual information; however, it incurs excessive computational cost when dealing with high-resolution inputs. In contrast, linear attention naturally enjoys linear complexity and has great potential to scale up to higher-resolution images. Nonetheless, the unsatisfactory performance of linear attention greatly limits its practical application in various scenarios. In this paper, we take a step forward to close the gap between the linear and Softmax attention with novel theoretical analyses, which demystify the core factors behind the performance deviations. Specifically, we present two key perspectives to understand and alleviate the limitations of linear attention: the injective property and the local modeling ability. Firstly, we prove that linear attention is not injective, which is prone to assign identical attention weights to different query vectors, thus adding to severe semantic confusion since different queries correspond to the same outputs. Secondly, we confirm that effective local modeling is essential for the success of Softmax attention, in which linear attention falls short. The aforementioned two fundamental differences significantly contribute to the disparities between these two attention paradigms, which is demonstrated by our substantial empirical validation in the paper. In addition, more experiment results indicate that linear attention, as long as endowed with these two properties, can outperform Softmax attention across various tasks while maintaining lower computation complexity. Code is available at https://github.com/LeapLabTHU/InLine.

Mamba is an effective state space model with linear computation complexity. It has recently shown impressive efficiency in dealing with high-resolution inputs across various vision tasks. In this paper, we reveal that the powerful Mamba model shares surprising similarities with linear attention Transformer, which typically underperform conventional Transformer in practice. By exploring the similarities and disparities between the effective Mamba and subpar linear attention Transformer, we provide comprehensive analyses to demystify the key factors behind Mamba’s success. Specifically, we reformulate the selective state space model and linear attention within a unified formulation, rephrasing Mamba as a variant of linear attention Transformer with six major distinctions: input gate, forget gate, shortcut, no attention normalization, single-head, and modified block design. For each design, we meticulously analyze its pros and cons, and empirically evaluate its impact on model performance in vision tasks. Interestingly, the results highlight the forget gate and block design as the core contributors to Mamba’s success, while the other four designs are less crucial. Based on these findings, we propose a Mamba- Inspired Linear Attention (MILA) model by incorporating the merits of these two key designs into linear attention. The resulting model outperforms various vision Mamba models in both image classification and high-resolution dense prediction tasks, while enjoying parallelizable computation and fast inference speed. Code is available at https://github.com/LeapLabTHU/MLLA.

Recently, token-based generation approaches have demonstrated their effectiveness in synthesizing visual content. As a representative example, non-autoregressive Transformers (NATs) can generate decent-quality images in just a few steps. NATs perform generation in a progressive manner, where the latent tokens of a resulting image are incrementally revealed step-by-step. At each step, the unrevealed image regions are padded with [MASK] tokens and inferred by NAT, with the most reliable predictions preserved as newly revealed, visible tokens. In this paper, we delve into understanding the mechanisms behind the effectiveness of NATs and uncover two important interaction patterns that naturally emerge from NAT’s paradigm: Spatially (within a step), although [MASK] and visible tokens are processed uniformly by NATs, the interactions between them are highly asymmetric. In specific, [MASK] tokens mainly gather information for decoding. On the contrary, visible tokens tend to primarily provide information, and their deep representations can be built only upon themselves. Temporally (across steps), the interactions between adjacent generation steps mostly concentrate on updating the representations of a few critical tokens, while the computation for the majority of tokens is generally repetitive. Driven by these findings, we propose EfficientNAT (ENAT), a NAT model that explicitly encourages these critical interactions inherent in NATs. At the spatial level, we disentangle the computations of visible and [MASK] tokens by encoding visible tokens independently, while decoding [MASK] tokens conditioned on the fully encoded visible tokens. At the temporal level, we prioritize the computation of the critical tokens at each step, while maximally reusing previously computed token representations to supplement necessary information. ENAT improves the performance of NATs notably with significantly reduced computational cost. Experiments on ImageNet-256 2 & 512 2 and MS-COCO validate the effectiveness of ENAT. Code and pre-trained models will be released at https://github.com/LeapLabTHU/ENAT.

Multimodal Large Language Models (MLLMs) have demonstrated remarkable comprehension and reasoning capabilities with complex language and visual data. These advances have spurred the vision of establishing a generalist robotic MLLM proficient in understanding complex human instructions and accomplishing various embodied tasks, whose feasibility has been recently verified~\cite{rt-2,rt-x}. However, developing MLLMs for real-world robots is challenging due to the typically limited computation and memory capacities available on robotic platforms. In contrast, the inference of MLLMs usually incorporates storing billions of parameters and performing tremendous computation, imposing significant hardware demands. In our paper, we seek to address this challenge by leveraging an intriguing observation: relatively easier situations make up the bulk of the procedure of controlling robots to fulfill diverse tasks, and they generally require far smaller models to obtain the correct robotic actions. Motivated by this observation, we propose a \emph{Dynamic Early-Exit for Robotic MLLM} (DeeR) framework that automatically adjusts the size of the activated MLLM based on each situation at hand. The approach leverages a multi-exit architecture in MLLMs, which allows the model to cease processing once a proper size of the model has been activated for a specific situation, thus avoiding further redundant computation. Additionally, we develop novel algorithms that establish early-termination criteria for DeeR, conditioned on predefined demands such as average computational cost (\emph{i.e.}, power consumption), as well as peak computational consumption (\emph{i.e.}, latency) and GPU memory usage. These enhancements ensure that DeeR operates efficiently under varying resource constraints while maintaining competitive performance. Moreover, we design a tailored training method for integrating temporal information on top of such multi-exit architectures to predict actions reasonably. On the CALVIN robot manipulation benchmark, DeeR demonstrates significant reductions in computational costs by 5.2-6.5x and GPU memory by 2x without compromising performance. Code and checkpoints are available at https://github.com/yueyang130/DeeR-VLA.

Existing parameter-efficient fine-tuning (PEFT) methods have achieved significant success on vision transformers (ViTs) adaptation by improving parameter efficiency. However, the exploration of enhancing inference efficiency during adaptation remains underexplored. This limits the broader application of pre-trained ViT models, especially when the model is computationally extensive. In this paper, we propose Dynamic Tuning (DyT), a novel approach to improve both parameter and inference efficiency for ViT adaptation. Specifically, besides using the lightweight adapter modules, we propose a token dispatcher to distinguish informative tokens from less important ones, allowing the latter to dynamically skip the original block, thereby reducing the redundant computation during inference. Additionally, we explore multiple design variants to find the best practice of DyT. Finally, inspired by the mixture-of-experts (MoE) mechanism, we introduce an enhanced adapter to further boost the adaptation performance. We validate DyT across various tasks, including image/video recognition and semantic segmentation. For instance, DyT achieves superior performance compared to existing PEFT methods while evoking only 71% of their FLOPs on the VTAB-1K benchmark.

Dynamic networks have become a pivotal area of study in deep learning due to their ability to selectively activate computing units (such as layers or channels) or dynamically allocate computation to information-rich regions. This capability significantly curtails unnecessary computations, adapting to varying inputs. Despite these advantages, the practical efficiency of dynamic models often falls short of theoretical computation. This discrepancy arises from three primary challenges: 1) a lack of a unified framework across different dynamic inference paradigms due to the fragmented research landscape; 2) an excessive focus on algorithm design at the expense of scheduling strategies, which are essential for optimizing resource utilization on hardware; and 3) the complexity of latency evaluation, since most current libraries cater to static operators. To tackle these issues, we introduce Latency-Aware Unified Dynamic Networks (LAUDNet), a general framework that integrates three fundamental dynamic paradigms-spatially-adaptive computation, layer skipping, and channel skipping-into a single unified formulation. LAUDNet not only refines algorithmic design but also enhances scheduling optimization with the aid of a latency predictor. This predictor efficiently and accurately predicts the inference latency of dynamic operators on specific hardware setups. Our empirical assessments across multiple vision tasks-image classification, object detection, and instance segmentation-confirm that LAUDNet significantly bridges the gap between theoretical and practical efficiency. For instance, LAUDNet cuts down the practical latency of its static counterpart, ResNet-101, by over 50% on hardware platforms like V100, RTX 3090, and TX2 GPUs. Additionally, LAUDNet excels in the accuracy-efficiency trade-off compared to other methods.

The superior performance of modern computer vision backbones (e.g., vision Transformers learned on ImageNet-1 K/22 K) usually comes with a costly training procedure. This study contributes to this issue by generalizing the idea of curriculum learning beyond its original formulation, i.e., training models using easier-to-harder data. Specifically, we reformulate the training curriculum as a soft-selection function, which uncovers progressively more difficult patterns within each example during training, instead of performing easier-to-harder sample selection. Our work is inspired by an intriguing observation on the learning dynamics of visual backbones: during the earlier stages of training, the model predominantly learns to recognize some 'easier-to-learn' discriminative patterns in the data. These patterns, when observed through frequency and spatial domains, incorporate lower-frequency components, and the natural image contents without distortion or data augmentation. Motivated by these findings, we propose a curriculum where the model always leverages all the training data at every learning stage, yet the exposure to the 'easier-to-learn' patterns of each example is initiated first, with harder patterns gradually introduced as training progresses. To implement this idea in a computationally efficient way, we introduce a cropping operation in the Fourier spectrum of the inputs, enabling the model to learn from only the lower-frequency components. Then we show that exposing the contents of natural images can be readily achieved by modulating the intensity of data augmentation. Finally, we integrate these two aspects and design curriculum learning schedules by proposing tailored searching algorithms. Moreover, we present useful techniques for deploying our approach efficiently in challenging practical scenarios, such as large-scale parallel training, and limited input/output or data pre-processing speed. The resulting method, EfficientTrain++, is simple, general, yet surprisingly effective. As an off-the-shelf approach, it reduces the training time of various popular models (e.g., ResNet, ConvNeXt, DeiT, PVT, Swin, CSWin, and CAFormer) by [Formula: see text] on ImageNet-1 K/22 K without sacrificing accuracy. It also demonstrates efficacy in self-supervised learning (e.g., MAE).

Differentiable architecture search (DARTS) has emerged as a promising technique for effective neural architecture search, and it mainly contains two steps to find the high-performance architecture. First, the DARTS supernet that consists of mixed operations will be optimized via gradient descent. Second, the final architecture will be built by the selected operations that contribute the most to the supernet. Although DARTS improves the efficiency of neural architecture search (NAS), it suffers from the well-known degeneration issue which can lead to deteriorating architectures. Existing works mainly attribute the degeneration issue to the failure of its supernet optimization, while little attention has been paid to the selection method. In this article, we cease to apply the widely-used magnitude-based selection method and propose a novel criterion based on operation strength that estimates the importance of an operation by its effect on the final loss. We show that the degeneration issue can be effectively addressed by using the proposed criterion without any modification of supernet optimization, indicating that the magnitude-based selection method can be a critical reason for the instability of DARTS. The experiments on NAS-Bench-201 and DARTS search spaces show the effectiveness of our method.

Fine-grained image recognition is a longstanding computer vision challenge that focuses on differentiating objects belonging to multiple subordinate categories within the same meta-category. Since images belonging to the same meta-category usually share similar visual appearances, mining discriminative visual cues is the key to distinguishing fine-grained categories. Although commonly used image-level data augmentation techniques have achieved great success in generic image classification problems, they are rarely applied in fine-grained scenarios, because their random editing-region behavior is prone to destroy the discriminative visual cues residing in the subtle regions. In this paper, we propose diversifying the training data at the feature-level to alleviate the discriminative region loss problem. Specifically, we produce diversified augmented samples by translating image features along semantically meaningful directions. The semantic directions are estimated with a covariance prediction network, which predicts a sample-wise covariance matrix to adapt to the large intra-class variation inherent in fine-grained images. Furthermore, the covariance prediction network is jointly optimized with the classification network in a meta-learning manner to alleviate the degenerate solution problem. Experiments on four competitive fine-grained recognition benchmarks (CUB-200-2011, Stanford Cars, FGVC Aircrafts, NABirds) demonstrate that our method significantly improves the generalization performance on several popular classification networks (e.g., ResNets, DenseNets, EfficientNets, RegNets and ViT). Combined with a recently proposed method, our semantic data augmentation approach achieves state-of-the-art performance on the CUB-200-2011 dataset. Source code is available at https://github.com/LeapLabTHU/LearnableISDA.

Early exiting has become a promising approach to improving the inference efficiency of deep networks. By structuring models with multiple classifiers (exits), predictions for "easy" samples can be generated at earlier exits, negating the need for executing deeper layers. Current multi-exit networks typically implement linear classifiers at intermediate layers, compelling low-level features to encapsulate high-level semantics. This sub-optimal design invariably undermines the performance of later exits. In this paper, we propose Dynamic Perceiver (Dyn-Perceiver) to decouple the feature extraction procedure and the early classification task with a novel dual-branch architecture. A feature branch serves to extract image features, while a classification branch processes a latent code assigned for classification tasks. Bi-directional cross-attention layers are established to progressively fuse the information of both branches. Early exits are placed exclusively within the classification branch, thus eliminating the need for linear separability in low-level features. Dyn-Perceiver constitutes a versatile and adaptable framework that can be built upon various architectures. Experiments on image classification, action recognition, and object detection demonstrate that our method significantly improves the inference efficiency of different backbones, outperforming numerous competitive approaches across a broad range of computational budgets. Evaluation on both CPU and GPU platforms substantiate the superior practical efficiency of Dyn-Perceiver. Code is available at https://www.github.com/LeapLabTHU/Dynamic_Perceiver.

The quadratic computation complexity of self-attention has been a persistent challenge when applying Transformer models to vision tasks. Linear attention, on the other hand, offers a much more efficient alternative with its linear complexity by approximating the Softmax operation through carefully designed mapping functions. However, current linear attention approaches either suffer from significant performance degradation or introduce additional computation overhead from the mapping functions. In this paper, we propose a novel Focused Linear Attention module to achieve both high efficiency and expressiveness. Specifically, we first analyze the factors contributing to the performance degradation of linear attention from two perspectives: the focus ability and feature diversity. To overcome these limitations, we introduce a simple yet effective mapping function and an efficient rank restoration module to enhance the expressiveness of self-attention while maintaining low computation complexity. Extensive experiments show that our linear attention module is applicable to a variety of advanced vision Transformers, and achieves consistently improved performances on multiple benchmarks. Code is available at https://github.com/LeapLabTHU/FLatten-Transformer.

Rotated object detection aims to identify and locate objects in images with arbitrary orientation. In this scenario, the oriented directions of objects vary considerably across different images, while multiple orientations of objects exist within an image. This intrinsic characteristic makes it challenging for standard backbone networks to extract high-quality features of these arbitrarily orientated objects. In this paper, we present Adaptive Rotated Convolution (ARC) module to handle the afore-mentioned challenges. In our ARC module, the convolution kernels rotate adaptively to extract object features with varying orientations in different images, and an efficient conditional computation mechanism is introduced to accommodate the large orientation variations of objects within an image. The two designs work seamlessly in rotated object detection problem. Moreover, ARC can conveniently serve as a plug-and-play module in various vision backbones to boost their representation ability to detect oriented objects accurately. Experiments on commonly used benchmarks (DOTA and HRSC2016) demonstrate that equipped with our proposed ARC module in the backbone network, the performance of multiple popular oriented object detectors is significantly improved (e.g. +3.03% mAP on Rotated RetinaNet and +4.16% on CFA). Combined with the highly competitive method Oriented R-CNN, the proposed approach achieves state-of-the-art performance on the DOTA dataset with 81.77% mAP. Code is available at https://github.com/LeapLabTHU/ARC.

Spatial-wise dynamic convolution has become a promising approach to improving the inference efficiency of deep networks. By allocating more computation to the most informative pixels, such an adaptive inference paradigm reduces the spatial redundancy in image features and saves a considerable amount of unnecessary computation. However, the theoretical efficiency achieved by previous methods can hardly translate into a realistic speedup, especially on the multi-core processors (e.g. GPUs). The key challenge is that the existing literature has only focused on designing algorithms with minimal computation, ignoring the fact that the practical latency can also be influenced by scheduling strategies and hardware properties. To bridge the gap between theoretical computation and practical efficiency, we propose a latency-aware spatial-wise dynamic network (LASNet), which performs coarse-grained spatially adaptive inference under the guidance of a novel latency prediction model. The latency prediction model can efficiently estimate the inference latency of dynamic networks by simultaneously considering algorithms, scheduling strategies, and hardware properties. We use the latency predictor to guide both the algorithm design and the scheduling optimization on various hardware platforms. Experiments on image classification, object detection and instance segmentation demonstrate that the proposed framework significantly improves the practical inference efficiency of deep networks. For example, the average latency of a ResNet-101 on the ImageNet validation set could be reduced by 36% and 46% on a server GPU (Nvidia Tesla-V100) and an edge device (Nvidia Jetson TX2 GPU) respectively without sacrificing the accuracy. Code is available at https://github.com/LeapLabTHU/LASNet.

Dynamic neural network is an emerging research topic in deep learning. Compared to static models which have fixed computational graphs and parameters at the inference stage, dynamic networks can adapt their structures or parameters to different inputs, leading to notable advantages in terms of accuracy, computational efficiency, adaptiveness, etc. In this survey, we comprehensively review this rapidly developing area by dividing dynamic networks into three main categories: 1) sample-wise dynamic models that process each sample with data-dependent architectures or parameters; 2) spatial-wise dynamic networks that conduct adaptive computation with respect to different spatial locations of image data; and 3) temporal-wise dynamic models that perform adaptive inference along the temporal dimension for sequential data such as videos and texts. The important research problems of dynamic networks, e.g., architecture design, decision making scheme, optimization technique and applications, are reviewed systematically. Finally, we discuss the open problems in this field together with interesting future research directions.

In this paper, we explore the spatial redundancy in video recognition with the aim to improve the computational efficiency. It is observed that the most informative region in each frame of a video is usually a small image patch, which shifts smoothly across frames. Therefore, we model the patch localization problem as a sequential decision task, and propose a reinforcement learning based approach for efficient spatially adaptive video recognition (AdaFocus). In specific, a light-weighted ConvNet is first adopted to quickly process the full video sequence, whose features are used by a recurrent policy network to localize the most task-relevant regions. Then the selected patches are inferred by a high-capacity network for the final prediction. During offline inference, once the informative patch sequence has been generated, the bulk of computation can be done in parallel, and is efficient on modern GPU devices. In addition, we demonstrate that the proposed method can be easily extended by further considering the temporal redundancy, e.g., dynamically skipping less valuable frames. Extensive experiments on five benchmark datasets, i.e., ActivityNet, FCVID, MiniKinetics, Something-Something V1&V2, demonstrate that our method is significantly more efficient than the competitive baselines. Code is available at https://github.com/blackfeather-wang/AdaFocus.

The backbone of traditional CNN classifier is generally considered as a feature extractor, followed by a linear layer which performs the classification. We propose a novel loss function, termed as CAM-loss, to constrain the embedded feature maps with the class activation maps (CAMs) which indicate the spatially discriminative regions of an image for particular categories. CAM-loss drives the backbone to express the features of target category and suppress the features of non-target categories or background, so as to obtain more discriminative feature representations. It can be simply applied in any CNN architecture with neglectable additional parameters and calculations. Experimental results show that CAM-loss is applicable to a variety of network structures and can be combined with mainstream regularization methods to improve the performance of image classification. The strong generalization ability of CAMloss is validated in the transfer learning and few shot learning tasks. Based on CAM-loss, we also propose a novel CAAM-CAM matching knowledge distillation method. This method directly uses the CAM generated by the teacher network to supervise the CAAM generated by the student network, which effectively improves the accuracy and convergence rate of the student network.

Spatial redundancy commonly exists in the learned representations of convolutional neural networks (CNNs), leading to unnecessary computation on high-resolution features. In this paper, we propose a novel Spatially Adaptive feature Refinement (SAR) approach to reduce such superfluous computation. It performs efficient inference by adaptively fusing information from two branches: one conducts standard convolution on input features at a lower spatial resolution, and the other one selectively refines a set of regions at the original resolution. The two branches complement each other in feature learning, and both of them evoke much less computation than standard convolution. SAR is a flexible method that can be conveniently plugged into existing CNNs to establish models with reduced spatial redundancy. Experiments on CIFAR and ImageNet classification, COCO object detection and PASCAL VOC semantic segmentation tasks validate that the proposed SAR can consistently improve the network performance and efficiency. Notably, our results show that SAR only refines less than 40% of the regions in the feature representations of a ResNet for 97% of the samples in the validation set of ImageNet to achieve comparable accuracy with the original model, revealing the high computational redundancy in the spatial dimension of CNNs.

Adaptive inference is an effective mechanism to achieve a dynamic tradeoff between accuracy and computational cost in deep networks. Existing works mainly exploit architecture redundancy in network depth or width. In this paper, we focus on spatial redundancy of input samples and propose a novel Resolution Adaptive Network (RANet), which is inspired by the intuition that low-resolution representations are sufficient for classifying “easy” inputs containing large objects with prototypical features, while only some “hard” samples need spatially detailed information. In RANet, the input images are first routed to a lightweight sub-network that efficiently extracts low-resolution representations, and those samples with high prediction confidence will exit early from the network without being further processed. Meanwhile, high-resolution paths in the network maintain the capability to recognize the “hard” samples. Therefore, RANet can effectively reduce the spatial redundancy involved in inferring high-resolution inputs. Empirically, we demonstrate the effectiveness of the proposed RANet on the CIFAR-10, CIFAR-100 and ImageNet datasets in both the anytime prediction setting and the budgeted batch classification setting.

FingerPing is a novel sensing technique that can recognize various fine-grained hand poses by analyzing acoustic resonance features. A surface-transducer mounted on a thumb ring injects acoustic chirps (20Hz to 6,000Hz) to the body. Four receivers distributed on the wrist and thumb collect the chirps. Different hand poses of the hand create distinct paths for the acoustic chirps to travel, creating unique frequency responses at the four receivers. We demonstrate how FingerPing can differentiate up to 22 hand poses, including the thumb touching each of the 12 phalanges on the hand as well as 10 American sign language poses. A user study with 16 participants showed that our system can recognize these two sets of poses with an accuracy of 93.77% and 95.64%, respectively. We discuss the opportunities and remaining challenges for the widespread use of this input technique.