Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling

We introduce InternVL 2.5, an advanced multimodal large language model (MLLM) series that builds upon InternVL 2.0, maintaining its core model architecture while introducing significant enhancements in training and testing strategies as well as data quality. In this work, we delve into the relationship between model scaling and performance, systematically exploring the performance trends in vision encoders, language models, dataset sizes, and test-time configurations. Through extensive evaluations on a wide range of benchmarks, including multi-discipline reasoning, document understanding, multi-image / video understanding, real-world comprehension, multimodal hallucination detection, visual grounding, multilingual capabilities, and pure language processing, InternVL 2.5 exhibits competitive performance, rivaling leading commercial models such as GPT-4o and Claude-3.5-Sonnet. Notably, our model is the first open-source MLLMs to surpass 70% on the MMMU benchmark, achieving a 3.7-point improvement through Chain-of-Thought (CoT) reasoning and showcasing strong potential for test-time scaling. We hope this model contributes to the open-source community by setting new standards for developing and applying multimodal AI systems. HuggingFace demo see https://huggingface.co/spaces/OpenGVLab/InternVL

GuardSplat: Efficient and Robust Watermarking for 3D Gaussian Splatting

3D Gaussian Splatting (3DGS) has recently created impressive assets for various applications. However, the copyright of these assets is not well protected as existing watermarking methods are not suited for 3DGS considering security, capacity, and invisibility. Besides, these methods often require hours or even days for optimization, limiting the application scenarios. In this paper, we propose GuardSplat, an innovative and efficient framework that effectively protects the copyright of 3DGS assets. Specifically, 1) We first propose a CLIP-guided Message Decoupling Optimization module for training the message decoder, leveraging CLIP's aligning capability and rich representations to achieve a high extraction accuracy with minimal optimization costs, presenting exceptional capability and efficiency. 2) Then, we propose a Spherical-harmonic-aware (SH-aware) Message Embedding module tailored for 3DGS, which employs a set of SH offsets to seamlessly embed the message into the SH features of each 3D Gaussian while maintaining the original 3D structure. It enables the 3DGS assets to be watermarked with minimal fidelity trade-offs and prevents malicious users from removing the messages from the model files, meeting the demands for invisibility and security. 3) We further propose an Anti-distortion Message Extraction module to improve robustness against various visual distortions. Extensive experiments demonstrate that GuardSplat outperforms the state-of-the-art methods and achieves fast optimization speed.

Learning Smooth Humanoid Locomotion through Lipschitz-Constrained Policies

Reinforcement learning combined with sim-to-real transfer offers a general framework for developing locomotion controllers for legged robots. To facilitate successful deployment in the real world, smoothing techniques, such as low-pass filters and smoothness rewards, are often employed to develop policies with smooth behaviors. However, because these techniques are non-differentiable and usually require tedious tuning of a large set of hyperparameters, they tend to require extensive manual tuning for each robotic platform. To address this challenge and establish a general technique for enforcing smooth behaviors, we propose a simple and effective method that imposes a Lipschitz constraint on a learned policy, which we refer to as Lipschitz-Constrained Policies (LCP). We show that the Lipschitz constraint can be implemented in the form of a gradient penalty, which provides a differentiable objective that can be easily incorporated with automatic differentiation frameworks. We demonstrate that LCP effectively replaces the need for smoothing rewards or low-pass filters and can be easily integrated into training frameworks for many distinct humanoid robots. We extensively evaluate LCP in both simulation and real-world humanoid robots, producing smooth and robust locomotion controllers. All simulation and deployment code, along with complete checkpoints, is available on our project page: https://lipschitz-constrained-policy.github.io.

Generalizable Humanoid Manipulation with Improved 3D Diffusion Policies

Humanoid robots capable of autonomous operation in diverse environments have long been a goal for roboticists. However, autonomous manipulation by humanoid robots has largely been restricted to one specific scene, primarily due to the difficulty of acquiring generalizable skills. Recent advances in 3D visuomotor policies, such as the 3D Diffusion Policy (DP3), have shown promise in extending these capabilities to wilder environments. However, 3D visuomotor policies often rely on camera calibration and point-cloud segmentation, which present challenges for deployment on mobile robots like humanoids. In this work, we introduce the Improved 3D Diffusion Policy (iDP3), a novel 3D visuomotor policy that eliminates these constraints by leveraging egocentric 3D visual representations. We demonstrate that iDP3 enables a full-sized humanoid robot to autonomously perform skills in diverse real-world scenarios, using only data collected in the lab. Videos are available at: https://humanoid-manipulation.github.io

GravMAD: Grounded Spatial Value Maps Guided Action Diffusion for Generalized 3D Manipulation

Robots' ability to follow language instructions and execute diverse 3D tasks is vital in robot learning. Traditional imitation learning-based methods perform well on seen tasks but struggle with novel, unseen ones due to variability. Recent approaches leverage large foundation models to assist in understanding novel tasks, thereby mitigating this issue. However, these methods lack a task-specific learning process, which is essential for an accurate understanding of 3D environments, often leading to execution failures. In this paper, we introduce GravMAD, a sub-goal-driven, language-conditioned action diffusion framework that combines the strengths of imitation learning and foundation models. Our approach breaks tasks into sub-goals based on language instructions, allowing auxiliary guidance during both training and inference. During training, we introduce Sub-goal Keypose Discovery to identify key sub-goals from demonstrations. Inference differs from training, as there are no demonstrations available, so we use pre-trained foundation models to bridge the gap and identify sub-goals for the current task. In both phases, GravMaps are generated from sub-goals, providing flexible 3D spatial guidance compared to fixed 3D positions. Empirical evaluations on RLBench show that GravMAD significantly outperforms state-of-the-art methods, with a 28.63% improvement on novel tasks and a 13.36% gain on tasks encountered during training. These results demonstrate GravMAD's strong multi-task learning and generalization in 3D manipulation. Video demonstrations are available at: https://gravmad.github.io.

Rina: Enhancing Ring-AllReduce with In-network Aggregation in Distributed Model Training

Parameter Server (PS) and Ring-AllReduce (RAR) are two widely utilized synchronization architectures in multi-worker Deep Learning (DL), also referred to as Distributed Deep Learning (DDL). However, PS encounters challenges with the ``incast'' issue, while RAR struggles with problems caused by the long dependency chain. The emerging In-network Aggregation (INA) has been proposed to integrate with PS to mitigate its incast issue. However, such PS-based INA has poor incremental deployment abilities as it requires replacing all the switches to show significant performance improvement, which is not cost-effective. In this study, we present the incorporation of INA capabilities into RAR, called RAR with In-Network Aggregation (Rina), to tackle both the problems above. Rina features its agent-worker mechanism. When an INA-capable ToR switch is deployed, all workers in this rack run as one abstracted worker with the help of the agent, resulting in both excellent incremental deployment capabilities and better throughput. We conducted extensive testbed and simulation evaluations to substantiate the throughput advantages of Rina over existing DDL training synchronization structures. Compared with the state-of-the-art PS-based INA methods ATP, Rina can achieve more than 50\% throughput with the same hardware cost.

VividDreamer: Towards High-Fidelity and Efficient Text-to-3D Generation

Text-to-3D generation aims to create 3D assets from text-to-image diffusion models. However, existing methods face an inherent bottleneck in generation quality because the widely-used objectives such as Score Distillation Sampling (SDS) inappropriately omit U-Net jacobians for swift generation, leading to significant bias compared to the "true" gradient obtained by full denoising sampling. This bias brings inconsistent updating direction, resulting in implausible 3D generation e.g., color deviation, Janus problem, and semantically inconsistent details). In this work, we propose Pose-dependent Consistency Distillation Sampling (PCDS), a novel yet efficient objective for diffusion-based 3D generation tasks. Specifically, PCDS builds the pose-dependent consistency function within diffusion trajectories, allowing to approximate true gradients through minimal sampling steps (1-3). Compared to SDS, PCDS can acquire a more accurate updating direction with the same sampling time (1 sampling step), while enabling few-step (2-3) sampling to trade compute for higher generation quality. For efficient generation, we propose a coarse-to-fine optimization strategy, which first utilizes 1-step PCDS to create the basic structure of 3D objects, and then gradually increases PCDS steps to generate fine-grained details. Extensive experiments demonstrate that our approach outperforms the state-of-the-art in generation quality and training efficiency, conspicuously alleviating the implausible 3D generation issues caused by the deviated updating direction. Moreover, it can be simply applied to many 3D generative applications to yield impressive 3D assets, please see our project page: https://narcissusex.github.io/VividDreamer.

CoCPF: Coordinate-based Continuous Projection Field for Ill-Posed Inverse Problem in Imaging

Sparse-view computed tomography (SVCT) reconstruction aims to acquire CT images based on sparsely-sampled measurements. It allows the subjects exposed to less ionizing radiation, reducing the lifetime risk of developing cancers. Recent researches employ implicit neural representation (INR) techniques to reconstruct CT images from a single SV sinogram. However, due to ill-posedness, these INR-based methods may leave considerable ``holes'' (i.e., unmodeled spaces) in their fields, leading to sub-optimal results. In this paper, we propose the Coordinate-based Continuous Projection Field (CoCPF), which aims to build hole-free representation fields for SVCT reconstruction, achieving better reconstruction quality. Specifically, to fill the holes, CoCPF first employs the stripe-based volume sampling module to broaden the sampling regions of Radon transformation from rays (1D space) to stripes (2D space), which can well cover the internal regions between SV projections. Then, by feeding the sampling regions into the proposed differentiable rendering modules, the holes can be jointly optimized during training, reducing the ill-posed levels. As a result, CoCPF can accurately estimate the internal measurements between SV projections (i.e., DV sinograms), producing high-quality CT images after re-projection. Extensive experiments on simulated and real projection datasets demonstrate that CoCPF outperforms state-of-the-art methods for 2D and 3D SVCT reconstructions under various projection numbers and geometries, yielding fine-grained details and fewer artifacts. Our code will be publicly available.

Visual Whole-Body Control for Legged Loco-Manipulation

We study the problem of mobile manipulation using legged robots equipped with an arm, namely legged loco-manipulation. The robot legs, while usually utilized for mobility, offer an opportunity to amplify the manipulation capabilities by conducting whole-body control. That is, the robot can control the legs and the arm at the same time to extend its workspace. We propose a framework that can conduct the whole-body control autonomously with visual observations. Our approach, namely Visual Whole-Body Control(VBC), is composed of a low-level policy using all degrees of freedom to track the end-effector manipulator position and a high-level policy proposing the end-effector position based on visual inputs. We train both levels of policies in simulation and perform Sim2Real transfer for real robot deployment. We perform extensive experiments and show significant improvements over baselines in picking up diverse objects in different configurations (heights, locations, orientations) and environments. Project page: https://wholebody-b1.github.io

HITSnDIFFs: From Truth Discovery to Ability Discovery by Recovering Matrices with the Consecutive Ones Property

We analyze a general problem in a crowd-sourced setting where one user asks a question (also called item) and other users return answers (also called labels) for this question. Different from existing crowd sourcing work which focuses on finding the most appropriate label for the question (the "truth"), our problem is to determine a ranking of the users based on their ability to answer questions. We call this problem "ability discovery" to emphasize the connection to and duality with the more well-studied problem of "truth discovery". To model items and their labels in a principled way, we draw upon Item Response Theory (IRT) which is the widely accepted theory behind standardized tests such as SAT and GRE. We start from an idealized setting where the relative performance of users is consistent across items and better users choose better fitting labels for each item. We posit that a principled algorithmic solution to our more general problem should solve this ideal setting correctly and observe that the response matrices in this setting obey the Consecutive Ones Property (C1P). While C1P is well understood algorithmically with various discrete algorithms, we devise a novel variant of the HITS algorithm which we call "HITSNDIFFS" (or HND), and prove that it can recover the ideal C1P-permutation in case it exists. Unlike fast combinatorial algorithms for finding the consecutive ones permutation (if it exists), HND also returns an ordering when such a permutation does not exist. Thus it provides a principled heuristic for our problem that is guaranteed to return the correct answer in the ideal setting. Our experiments show that HND produces user rankings with robustly high accuracy compared to state-of-the-art truth discovery methods. We also show that our novel variant of HITS scales better in the number of users than ABH, the only prior spectral C1P reconstruction algorithm.