Inference
-
Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis
2024-12-04 18:59:24Qitao Zhao, Shubham TulsianiAbstract
Inferring the 3D structure underlying a set of multi-view images typically requires solving two co-dependent tasks -- accurate 3D reconstruction requires precise camera poses, and predicting camera poses relies on (implicitly or explicitly) modeling the underlying 3D. The classical framework of analysis by synthesis casts this inference as a joint optimization seeking to explain the observed pixels, and recent instantiations learn expressive 3D representations (e.g., Neural Fields) with gradient-descent-based pose refinement of initial pose estimates. However, given a sparse set of observed views, the observations may not provide sufficient direct evidence to obtain complete and accurate 3D. Moreover, large errors in pose estimation may not be easily corrected and can further degrade the inferred 3D. To allow robust 3D reconstruction and pose estimation in this challenging setup, we propose SparseAGS, a method that adapts this analysis-by-synthesis approach by: a) including novel-view-synthesis-based generative priors in conjunction with photometric objectives to improve the quality of the inferred 3D, and b) explicitly reasoning about outliers and using a discrete search with a continuous optimization-based strategy to correct them. We validate our framework across real-world and synthetic datasets in combination with several off-the-shelf pose estimation systems as initialization. We find that it significantly improves the base systems' pose accuracy while yielding high-quality 3D reconstructions that outperform the results from current multi-view reconstruction baselines.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03570
PDF
-
Automated Test-Case Generation for REST APIs Using Model Inference Search Heuristic
Abstract
The rising popularity of the microservice architectural style has led to a growing demand for automated testing approaches tailored to these systems. EvoMaster is a state-of-the-art tool that uses Evolutionary Algorithms (EAs) to automatically generate test cases for microservices' REST APIs. One limitation of these EAs is the use of unit-level search heuristics, such as branch distances, which focus on fine-grained code coverage and may not effectively capture the complex, interconnected behaviors characteristic of system-level testing. To address this limitation, we propose a new search heuristic (MISH) that uses real-time automaton learning to guide the test case generation process. We capture the sequential call patterns exhibited by a test case by learning an automaton from the stream of log events outputted by different microservices within the same system. Therefore, MISH learns a representation of the systemwide behavior, allowing us to define the fitness of a test case based on the path it traverses within the inferred automaton. We empirically evaluate MISH's effectiveness on six real-world benchmark microservice applications and compare it against a state-of-the-art technique, MOSA, for testing REST APIs. Our evaluation shows promising results for using MISH to guide the automated test case generation within EvoMaster.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03420
PDF
-
PrefixKV: Adaptive Prefix KV Cache is What Vision Instruction-Following Models Need for Efficient Generation
2024-12-04 15:48:59Ao Wang, Hui Chen, Jianchao Tan, Kefeng Zhang, Xunliang Cai, Zijia Lin, Jungong Han, Guiguang DingAbstract
Recently, large vision-language models (LVLMs) have rapidly gained popularity for their strong generation and reasoning capabilities given diverse multimodal inputs. However, these models incur significant computational and memory overhead during inference, which greatly hinders the efficient deployment in practical scenarios. The extensive key-value (KV) cache, necessitated by the lengthy input and output sequences, notably contributes to the high inference cost. Based on this, recent works have investigated ways to reduce the KV cache size for higher efficiency. Although effective, they generally overlook the distinct importance distributions of KV vectors across layers and maintain the same cache size for each layer during the next token prediction. This results in the significant contextual information loss for certain layers, leading to notable performance decline. To address this, we present PrefixKV. It reframes the challenge of determining KV cache sizes for all layers into the task of searching for the optimal global prefix configuration. With an adaptive layer-wise KV retention recipe based on binary search, the maximum contextual information can thus be preserved in each layer, facilitating the generation. Extensive experiments demonstrate that our method achieves the state-of-the-art performance compared with others. It exhibits superior inference efficiency and generation quality trade-offs, showing promising potential for practical applications. Code is available at \url{this https URL}.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03409
PDF
-
Volumetrically Consistent 3D Gaussian Rasterization
2024-12-04 15:05:43Chinmay Talegaonkar, Yash Belhe, Ravi Ramamoorthi, Nicholas AntipaAbstract
Recently, 3D Gaussian Splatting (3DGS) has enabled photorealistic view synthesis at high inference speeds. However, its splatting-based rendering model makes several approximations to the rendering equation, reducing physical accuracy. We show that splatting and its approximations are unnecessary, even within a rasterizer; we instead volumetrically integrate 3D Gaussians directly to compute the transmittance across them analytically. We use this analytic transmittance to derive more physically-accurate alpha values than 3DGS, which can directly be used within their framework. The result is a method that more closely follows the volume rendering equation (similar to ray-tracing) while enjoying the speed benefits of rasterization. Our method represents opaque surfaces with higher accuracy and fewer points than 3DGS. This enables it to outperform 3DGS for view synthesis (measured in SSIM and LPIPS). Being volumetrically consistent also enables our method to work out of the box for tomography. We match the state-of-the-art 3DGS-based tomography method with fewer points. Being volumetrically consistent also enables our method to work out of the box for tomography. We match the state-of-the-art 3DGS-based tomography method with fewer points.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03378
PDF
-
A Stitch in Time Saves Nine: Small VLM is a Precise Guidance for accelerating Large VLMs
2024-12-04 13:56:44Wangbo Zhao, Yizeng Han, Jiasheng Tang, Zhikai Li, Yibing Song, Kai Wang, Zhangyang Wang, Yang YouAbstract
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 requires 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 \textbf{training-free} method, \underline{\textbf{S}}mall VLM \underline{\textbf{G}}uidance for accelerating \underline{\textbf{L}}arge VLMs (\textbf{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.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03324
PDF
-
Path-Guided Particle-based Sampling
2024-12-04 13:44:56Mingzhou Fan, Ruida Zhou, Chao Tian, Xiaoning QianAbstract
Particle-based Bayesian inference methods by sampling from a partition-free target (posterior) distribution, e.g., Stein variational gradient descent (SVGD), have attracted significant attention. We propose a path-guided particle-based sampling~(PGPS) method based on a novel Log-weighted Shrinkage (LwS) density path linking an initial distribution to the target distribution. We propose to utilize a Neural network to learn a vector field motivated by the Fokker-Planck equation of the designed density path. Particles, initiated from the initial distribution, evolve according to the ordinary differential equation defined by the vector field. The distribution of these particles is guided along a density path from the initial distribution to the target distribution. The proposed LwS density path allows for an efficient search of modes of the target distribution while canonical methods fail. We theoretically analyze the Wasserstein distance of the distribution of the PGPS-generated samples and the target distribution due to approximation and discretization errors. Practically, the proposed PGPS-LwS method demonstrates higher Bayesian inference accuracy and better calibration ability in experiments conducted on both synthetic and real-world Bayesian learning tasks, compared to baselines, such as SVGD and Langevin dynamics, etc.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03312
PDF
-
Diffusion-VLA: Scaling Robot Foundation Models via Unified Diffusion and Autoregression
2024-12-04 13:11:38Junjie Wen, Minjie Zhu, Yichen Zhu, Zhibin Tang, Jinming Li, Zhongyi Zhou, Chengmeng Li, Xiaoyu Liu, Yaxin Peng, Chaomin Shen, Feifei FengAbstract
In this paper, we present DiffusionVLA, a novel framework that seamlessly combines the autoregression model with the diffusion model for learning visuomotor policy. Central to our approach is a next-token prediction objective, enabling the model to reason effectively over the user's query in the context of current observations. Subsequently, a diffusion model is attached to generate robust action outputs. To enhance policy learning through self-reasoning, we introduce a novel reasoning injection module that integrates reasoning phrases directly into the policy learning process. The whole framework is simple and flexible, making it easy to deploy and upgrade. We conduct extensive experiments using multiple real robots to validate the effectiveness of DiffusionVLA. Our tests include a challenging factory sorting task, where DiffusionVLA successfully categorizes objects, including those not seen during training. We observe that the reasoning module makes the model interpretable. It allows observers to understand the model thought process and identify potential causes of policy failures. Additionally, we test DiffusionVLA on a zero-shot bin-picking task, achieving 63.7\% accuracy on 102 previously unseen objects. Our method demonstrates robustness to visual changes, such as distractors and new backgrounds, and easily adapts to new embodiments. Furthermore, DiffusionVLA can follow novel instructions and retain conversational ability. Notably, DiffusionVLA is data-efficient and fast at inference; our smallest DiffusionVLA-2B runs 82Hz on a single A6000 GPU and can train from scratch on less than 50 demonstrations for a complex task. Finally, we scale the model from 2B to 72B parameters, showcasing improved generalization capabilities with increased model size.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03293
PDF
-
Detecting abnormal heart sound using mobile phones and on-device IConNet
Abstract
Given the global prevalence of cardiovascular diseases, there is a pressing need for easily accessible early screening methods. Typically, this requires medical practitioners to investigate heart auscultations for irregular sounds, followed by echocardiography and electrocardiography tests. To democratize early diagnosis, we present a user-friendly solution for abnormal heart sound detection, utilizing mobile phones and a lightweight neural network optimized for on-device inference. Unlike previous approaches reliant on specialized stethoscopes, our method directly analyzes audio recordings, facilitated by a novel architecture known as IConNet. IConNet, an Interpretable Convolutional Neural Network, harnesses insights from audio signal processing, enhancing efficiency and providing transparency in neural pattern extraction from raw waveform signals. This is a significant step towards trustworthy AI in healthcare, aiding in remote health monitoring efforts.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03267
PDF
-
AIM: Adaptive Inference of Multi-Modal LLMs via Token Merging and Pruning
2024-12-04 11:47:57Yiwu Zhong, Zhuoming Liu, Yin Li, Liwei WangAbstract
Large language models (LLMs) have enabled the creation of multi-modal LLMs that exhibit strong comprehension of visual data such as images and videos. However, these models usually rely on extensive visual tokens from visual encoders, leading to high computational demands, which limits their applicability in resource-constrained environments and for long-context tasks. In this work, we propose a training-free adaptive inference method for multi-modal LLMs that can accommodate a broad range of efficiency requirements with a minimum performance drop. Our method consists of a) iterative token merging based on embedding similarity before LLMs, and b) progressive token pruning within LLM layers based on multi-modal importance. With a minimalist design, our method can be applied to both video and image LLMs. Extensive experiments on diverse video and image benchmarks demonstrate that, our method substantially reduces computation load (e.g., a $\textbf{7-fold}$ reduction in FLOPs) while preserving the performance of video and image LLMs. Further, under a similar computational cost, our method outperforms the state-of-the-art methods in long video understanding (e.g., $\textbf{+4.6}$ on MLVU). Additionally, our in-depth analysis provides insights into token redundancy and LLM layer behaviors, offering guidance for future research in designing efficient multi-modal LLMs. Our code will be available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03248
PDF
-
Continual Low-Rank Scaled Dot-product Attention
2024-12-04 11:05:01Gin\'es Carreto Pic\'on, Illia Oleksiienko, Lukas Hedegaard, Arian Bakhtiarnia, Alexandros IosifidisAbstract
Transformers are widely used for their ability to capture data relations in sequence processing, with great success for a wide range of static tasks. However, the computational and memory footprint of their main component, i.e., the Scaled Dot-product Attention, is commonly overlooked. This makes their adoption in applications involving stream data processing with constraints in response latency, computational and memory resources infeasible. Some works have proposed methods to lower the computational cost of transformers, i.e. low-rank approximations, sparsity in attention, and efficient formulations for Continual Inference. In this paper, we introduce a new formulation of the Scaled Dot-product Attention based on the Nyström approximation that is suitable for Continual Inference. In experiments on Online Audio Classification and Online Action Detection tasks, the proposed Continual Scaled Dot-product Attention can lower the number of operations by up to three orders of magnitude compared to the original Transformers while retaining the predictive performance of competing models.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03214
PDF
-
ClusterKV: Manipulating LLM KV Cache in Semantic Space for Recallable Compression
2024-12-04 10:58:27Guangda Liu, Chengwei Li, Jieru Zhao, Chenqi Zhang, Minyi GuoAbstract
Large Language Models (LLMs) have been widely deployed in a variety of applications, and the context length is rapidly increasing to handle tasks such as long-document QA and complex logical reasoning. However, long context poses significant challenges for inference efficiency, including high memory costs of key-value (KV) cache and increased latency due to extensive memory accesses. Recent works have proposed compressing KV cache to approximate computation, but these methods either evict tokens permanently, never recalling them for later inference, or recall previous tokens at the granularity of pages divided by textual positions. Both approaches degrade the model accuracy and output quality. To achieve efficient and accurate recallable KV cache compression, we introduce ClusterKV, which recalls tokens at the granularity of semantic clusters. We design and implement efficient algorithms and systems for clustering, selection, indexing and caching. Experiment results show that ClusterKV attains negligible accuracy loss across various tasks with 32k context lengths, using only a 1k to 2k KV cache budget, and achieves up to a 2$\times$ speedup in latency and a 2.5$\times$ improvement in decoding throughput. Compared to SoTA recallable KV compression methods, ClusterKV demonstrates higher model accuracy and output quality, while maintaining or exceeding inference efficiency.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03213
PDF
-
Optimizing Dense Visual Predictions Through Multi-Task Coherence and Prioritization
2024-12-04 10:05:47Maxime Fontana, Michael Spratling, Miaojing ShiAbstract
Multi-Task Learning (MTL) involves the concurrent training of multiple tasks, offering notable advantages for dense prediction tasks in computer vision. MTL not only reduces training and inference time as opposed to having multiple single-task models, but also enhances task accuracy through the interaction of multiple tasks. However, existing methods face limitations. They often rely on suboptimal cross-task interactions, resulting in task-specific predictions with poor geometric and predictive coherence. In addition, many approaches use inadequate loss weighting strategies, which do not address the inherent variability in task evolution during training. To overcome these challenges, we propose an advanced MTL model specifically designed for dense vision tasks. Our model leverages state-of-the-art vision transformers with task-specific decoders. To enhance cross-task coherence, we introduce a trace-back method that improves both cross-task geometric and predictive features. Furthermore, we present a novel dynamic task balancing approach that projects task losses onto a common scale and prioritizes more challenging tasks during training. Extensive experiments demonstrate the superiority of our method, establishing new state-of-the-art performance across two benchmark datasets. The code is available at:this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2412.03179
PDF
-
Appearance Matching Adapter for Exemplar-based Semantic Image Synthesis
2024-12-04 09:17:47Siyoon Jin, Jisu Nam, Jiyoung Kim, Dahyun Chung, Yeong-Seok Kim, Joonhyung Park, Heonjeong Chu, Seungryong KimAbstract
Exemplar-based semantic image synthesis aims to generate images aligned with given semantic content while preserving the appearance of an exemplar image. Conventional structure-guidance models, such as ControlNet, are limited in that they cannot directly utilize exemplar images as input, relying instead solely on text prompts to control appearance. Recent tuning-free approaches address this limitation by transferring local appearance from the exemplar image to the synthesized image through implicit cross-image matching in the augmented self-attention mechanism of pre-trained diffusion models. However, these methods face challenges when applied to content-rich scenes with significant geometric deformations, such as driving scenes. In this paper, we propose the Appearance Matching Adapter (AM-Adapter), a learnable framework that enhances cross-image matching within augmented self-attention by incorporating semantic information from segmentation maps. To effectively disentangle generation and matching processes, we adopt a stage-wise training approach. Initially, we train the structure-guidance and generation networks, followed by training the AM-Adapter while keeping the other networks frozen. During inference, we introduce an automated exemplar retrieval method to efficiently select exemplar image-segmentation pairs. Despite utilizing a limited number of learnable parameters, our method achieves state-of-the-art performance, excelling in both semantic alignment preservation and local appearance fidelity. Extensive ablation studies further validate our design choices. Code and pre-trained weights will be publicly available.: this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2412.03150
PDF
-
Lightweight Multiplane Images Network for Real-Time Stereoscopic Conversion from Planar Video
2024-12-04 08:04:14Shanding Diao, Yang Zhao, Yuan Chen, Zhao Zhang, Wei Jia, Ronggang WangAbstract
With the rapid development of stereoscopic display technologies, especially glasses-free 3D screens, and virtual reality devices, stereoscopic conversion has become an important task to address the lack of high-quality stereoscopic image and video resources. Current stereoscopic conversion algorithms typically struggle to balance reconstruction performance and inference efficiency. This paper proposes a planar video real-time stereoscopic conversion network based on multi-plane images (MPI), which consists of a detail branch for generating MPI and a depth-semantic branch for perceiving depth information. Unlike models that depend on explicit depth map inputs, the proposed method employs a lightweight depth-semantic branch to extract depth-aware features implicitly. To optimize the lightweight branch, a heavy training but light inference strategy is adopted, which involves designing a coarse-to-fine auxiliary branch that is only used during the training stage. In addition, the proposed method simplifies the MPI rendering process for stereoscopic conversion scenarios to further accelerate the inference. Experimental results demonstrate that the proposed method can achieve comparable performance to some state-of-the-art (SOTA) models and support real-time inference at 2K resolution. Compared to the SOTA TMPI algorithm, the proposed method obtains similar subjective quality while achieving over $40\times$ inference acceleration.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03102
PDF
-
Real-Time AIoT for UAV Antenna Interference Detection via Edge-Cloud Collaboration
2024-12-04 06:20:36Jun Dong, Jintao Cheng, Jin Wu, Chengxi Zhang, Shunyi Zhao, Xiaoyu TangAbstract
In the fifth-generation (5G) era, eliminating communication interference sources is crucial for maintaining network performance. Interference often originates from unauthorized or malfunctioning antennas, and radio monitoring agencies must address numerous sources of such antennas annually. Unmanned aerial vehicles (UAVs) can improve inspection efficiency. However, the data transmission delay in the existing cloud-only (CO) artificial intelligence (AI) mode fails to meet the low latency requirements for real-time performance. Therefore, we propose a computer vision-based AI of Things (AIoT) system to detect antenna interference sources for UAVs. The system adopts an optimized edge-cloud collaboration (ECC+) mode, combining a keyframe selection algorithm (KSA), focusing on reducing end-to-end latency (E2EL) and ensuring reliable data transmission, which aligns with the core principles of ultra-reliable low-latency communication (URLLC). At the core of our approach is an end-to-end antenna localization scheme based on the tracking-by-detection (TBD) paradigm, including a detector (EdgeAnt) and a tracker (AntSort). EdgeAnt achieves state-of-the-art (SOTA) performance with a mean average precision (mAP) of 42.1% on our custom antenna interference source dataset, requiring only 3 million parameters and 14.7 GFLOPs. On the COCO dataset, EdgeAnt achieves 38.9% mAP with 5.4 GFLOPs. We deployed EdgeAnt on Jetson Xavier NX (TRT) and Raspberry Pi 4B (NCNN), achieving real-time inference speeds of 21.1 (1088) and 4.8 (640) frames per second (FPS), respectively. Compared with CO mode, the ECC+ mode reduces E2EL by 88.9%, increases accuracy by 28.2%. Additionally, the system offers excellent scalability for coordinated multiple UAVs inspections. The detector code is publicly available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03055
PDF
-
Frequency-Guided Diffusion Model with Perturbation Training for Skeleton-Based Video Anomaly Detection
2024-12-04 05:43:53Xiaofeng Tan, Hongsong Wang, Xin GengAbstract
Video anomaly detection is an essential yet challenging open-set task in computer vision, often addressed by leveraging reconstruction as a proxy task. However, existing reconstruction-based methods encounter challenges in two main aspects: (1) limited model robustness for open-set scenarios, (2) and an overemphasis on, but restricted capacity for, detailed motion reconstruction. To this end, we propose a novel frequency-guided diffusion model with perturbation training, which enhances the model robustness by perturbation training and emphasizes the principal motion components guided by motion frequencies. Specifically, we first use a trainable generator to produce perturbative samples for perturbation training of the diffusion model. During the perturbation training phase, the model robustness is enhanced and the domain of the reconstructed model is broadened by training against this generator. Subsequently, perturbative samples are introduced for inference, which impacts the reconstruction of normal and abnormal motions differentially, thereby enhancing their separability. Considering that motion details originate from high-frequency information, we propose a masking method based on 2D discrete cosine transform to separate high-frequency information and low-frequency information. Guided by the high-frequency information from observed motion, the diffusion model can focus on generating low-frequency information, and thus reconstructing the motion accurately. Experimental results on five video anomaly detection datasets, including human-related and open-set benchmarks, demonstrate the effectiveness of the proposed method. Our code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03044
PDF
-
Pixel-level and Semantic-level Adjustable Super-resolution: A Dual-LoRA Approach
2024-12-04 04:07:49Lingchen Sun, Rongyuan Wu, Zhiyuan Ma, Shuaizheng Liu, Qiaosi Yi, Lei ZhangAbstract
Diffusion prior-based methods have shown impressive results in real-world image super-resolution (SR). However, most existing methods entangle pixel-level and semantic-level SR objectives in the training process, struggling to balance pixel-wise fidelity and perceptual quality. Meanwhile, users have varying preferences on SR results, thus it is demanded to develop an adjustable SR model that can be tailored to different fidelity-perception preferences during inference without re-training. We present Pixel-level and Semantic-level Adjustable SR (PiSA-SR), which learns two LoRA modules upon the pre-trained stable-diffusion (SD) model to achieve improved and adjustable SR results. We first formulate the SD-based SR problem as learning the residual between the low-quality input and the high-quality output, then show that the learning objective can be decoupled into two distinct LoRA weight spaces: one is characterized by the $\ell_2$-loss for pixel-level regression, and another is characterized by the LPIPS and classifier score distillation losses to extract semantic information from pre-trained classification and SD models. In its default setting, PiSA-SR can be performed in a single diffusion step, achieving leading real-world SR results in both quality and efficiency. By introducing two adjustable guidance scales on the two LoRA modules to control the strengths of pixel-wise fidelity and semantic-level details during inference, PiSASR can offer flexible SR results according to user preference without re-training. Codes and models can be found at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2412.03017
PDF
-
Partially Conditioned Patch Parallelism for Accelerated Diffusion Model Inference
2024-12-04 02:12:50XiuYu Zhang, Zening Luo, Michelle E. LuAbstract
Diffusion models have exhibited exciting capabilities in generating images and are also very promising for video creation. However, the inference speed of diffusion models is limited by the slow sampling process, restricting its use cases. The sequential denoising steps required for generating a single sample could take tens or hundreds of iterations and thus have become a significant bottleneck. This limitation is more salient for applications that are interactive in nature or require small latency. To address this challenge, we propose Partially Conditioned Patch Parallelism (PCPP) to accelerate the inference of high-resolution diffusion models. Using the fact that the difference between the images in adjacent diffusion steps is nearly zero, Patch Parallelism (PP) leverages multiple GPUs communicating asynchronously to compute patches of an image in multiple computing devices based on the entire image (all patches) in the previous diffusion step. PCPP develops PP to reduce computation in inference by conditioning only on parts of the neighboring patches in each diffusion step, which also decreases communication among computing devices. As a result, PCPP decreases the communication cost by around $70\%$ compared to DistriFusion (the state of the art implementation of PP) and achieves $2.36\sim 8.02\times$ inference speed-up using $4\sim 8$ GPUs compared to $2.32\sim 6.71\times$ achieved by DistriFusion depending on the computing device configuration and resolution of generation at the cost of a possible decrease in image quality. PCPP demonstrates the potential to strike a favorable trade-off, enabling high-quality image generation with substantially reduced latency.
Abstract (translated)
URL
https://arxiv.org/abs/2412.02962
PDF
-
EgoCast: Forecasting Egocentric Human Pose in the Wild
2024-12-03 23:11:06Maria Escobar, Juanita Puentes, Cristhian Forigua, Jordi Pont-Tuset, Kevis-Kokitsi Maninis, Pablo ArbelaezAbstract
Accurately estimating and forecasting human body pose is important for enhancing the user's sense of immersion in Augmented Reality. Addressing this need, our paper introduces EgoCast, a bimodal method for 3D human pose forecasting using egocentric videos and proprioceptive data. We study the task of human pose forecasting in a realistic setting, extending the boundaries of temporal forecasting in dynamic scenes and building on the current framework for current pose estimation in the wild. We introduce a current-frame estimation module that generates pseudo-groundtruth poses for inference, eliminating the need for past groundtruth poses typically required by current methods during forecasting. Our experimental results on the recent Ego-Exo4D and Aria Digital Twin datasets validate EgoCast for real-life motion estimation. On the Ego-Exo4D Body Pose 2024 Challenge, our method significantly outperforms the state-of-the-art approaches, laying the groundwork for future research in human pose estimation and forecasting in unscripted activities with egocentric inputs.
Abstract (translated)
URL
https://arxiv.org/abs/2412.02903
PDF
-
Effortless Efficiency: Low-Cost Pruning of Diffusion Models
2024-12-03 21:37:50Yang Zhang, Er Jin, Yanfei Dong, Ashkan Khakzar, Philip Torr, Johannes Stegmaier, Kenji KawaguchiAbstract
Diffusion models have achieved impressive advancements in various vision tasks. However, these gains often rely on increasing model size, which escalates computational complexity and memory demands, complicating deployment, raising inference costs, and causing environmental impact. While some studies have explored pruning techniques to improve the memory efficiency of diffusion models, most existing methods require extensive retraining to retain the model performance. Retraining a modern large diffusion model is extremely costly and resource-intensive, which limits the practicality of these methods. In this work, we achieve low-cost diffusion pruning without retraining by proposing a model-agnostic structural pruning framework for diffusion models that learns a differentiable mask to sparsify the model. To ensure effective pruning that preserves the quality of the final denoised latent, we design a novel end-to-end pruning objective that spans the entire diffusion process. As end-to-end pruning is memory-intensive, we further propose time step gradient checkpointing, a technique that significantly reduces memory usage during optimization, enabling end-to-end pruning within a limited memory budget. Results on state-of-the-art U-Net diffusion models SDXL and diffusion transformers (FLUX) demonstrate that our method can effectively prune up to 20% parameters with minimal perceptible performance degradation, and notably, without the need for model retraining. We also showcase that our method can still prune on top of time step distilled diffusion models.
Abstract (translated)
URL
https://arxiv.org/abs/2412.02852
PDF
