Guo Chen

Amala Sanjay Deshmukh, Kateryna Chumachenko, Tuomas Rintamaki et al. · nvidia

We introduce Nemotron Nano V2 VL, the latest model of the Nemotron vision-language series designed for strong real-world document understanding, long video comprehension, and reasoning tasks. Nemotron Nano V2 VL delivers significant improvements over our previous model, Llama-3.1-Nemotron-Nano-VL-8B, across all vision and text domains through major enhancements in model architecture, datasets, and training recipes. Nemotron Nano V2 VL builds on Nemotron Nano V2, a hybrid Mamba-Transformer LLM, and innovative token reduction techniques to achieve higher inference throughput in long document and video scenarios. We are releasing model checkpoints in BF16, FP8, and FP4 formats and sharing large parts of our datasets, recipes and training code.

Aaron Blakeman, Aarti Basant, Abhinav Khattar et al. · nvidia

As inference-time scaling becomes critical for enhanced reasoning capabilities, it is increasingly becoming important to build models that are efficient to infer. We introduce Nemotron-H, a family of 8B and 56B/47B hybrid Mamba-Transformer models designed to reduce inference cost for a given accuracy level. To achieve this goal, we replace the majority of self-attention layers in the common Transformer model architecture with Mamba layers that perform constant computation and require constant memory per generated token. We show that Nemotron-H models offer either better or on-par accuracy compared to other similarly-sized state-of-the-art open-sourced Transformer models (e.g., Qwen-2.5-7B/72B and Llama-3.1-8B/70B), while being up to 3$\times$ faster at inference. To further increase inference speed and reduce the memory required at inference time, we created Nemotron-H-47B-Base from the 56B model using a new compression via pruning and distillation technique called MiniPuzzle. Nemotron-H-47B-Base achieves similar accuracy to the 56B model, but is 20% faster to infer. In addition, we introduce an FP8-based training recipe and show that it can achieve on par results with BF16-based training. This recipe is used to train the 56B model. We are releasing Nemotron-H base model checkpoints with support in Hugging Face and NeMo.

Yi Wang, Kunchang Li, Yizhuo Li et al.

The foundation models have recently shown excellent performance on a variety of downstream tasks in computer vision. However, most existing vision foundation models simply focus on image-level pretraining and adpation, which are limited for dynamic and complex video-level understanding tasks. To fill the gap, we present general video foundation models, InternVideo, by taking advantage of both generative and discriminative self-supervised video learning. Specifically, InternVideo efficiently explores masked video modeling and video-language contrastive learning as the pretraining objectives, and selectively coordinates video representations of these two complementary frameworks in a learnable manner to boost various video applications. Without bells and whistles, InternVideo achieves state-of-the-art performance on 39 video datasets from extensive tasks including video action recognition/detection, video-language alignment, and open-world video applications. Especially, our methods can obtain 91.1% and 77.2% top-1 accuracy on the challenging Kinetics-400 and Something-Something V2 benchmarks, respectively. All of these results effectively show the generality of our InternVideo for video understanding. The code will be released at https://github.com/OpenGVLab/InternVideo .

Kunchang Li, Yali Wang, Yinan He et al.

With the rapid development of Multi-modal Large Language Models (MLLMs), a number of diagnostic benchmarks have recently emerged to evaluate the comprehension capabilities of these models. However, most benchmarks predominantly assess spatial understanding in the static image tasks, while overlooking temporal understanding in the dynamic video tasks. To alleviate this issue, we introduce a comprehensive Multi-modal Video understanding Benchmark, namely MVBench, which covers 20 challenging video tasks that cannot be effectively solved with a single frame. Specifically, we first introduce a novel static-to-dynamic method to define these temporal-related tasks. By transforming various static tasks into dynamic ones, we enable the systematic generation of video tasks that require a broad spectrum of temporal skills, ranging from perception to cognition. Then, guided by the task definition, we automatically convert public video annotations into multiple-choice QA to evaluate each task. On one hand, such a distinct paradigm allows us to build MVBench efficiently, without much manual intervention. On the other hand, it guarantees evaluation fairness with ground-truth video annotations, avoiding the biased scoring of LLMs. Moreover, we further develop a robust video MLLM baseline, i.e., VideoChat2, by progressive multi-modal training with diverse instruction-tuning data. The extensive results on our MVBench reveal that, the existing MLLMs are far from satisfactory in temporal understanding, while our VideoChat2 largely surpasses these leading models by over 15% on MVBench. All models and data are available at https://github.com/OpenGVLab/Ask-Anything.

Min Yang, Guo Chen, Yin-Dong Zheng et al.

Temporal action detection (TAD) is extensively studied in the video understanding community by generally following the object detection pipeline in images. However, complex designs are not uncommon in TAD, such as two-stream feature extraction, multi-stage training, complex temporal modeling, and global context fusion. In this paper, we do not aim to introduce any novel technique for TAD. Instead, we study a simple, straightforward, yet must-known baseline given the current status of complex design and low detection efficiency in TAD. In our simple baseline (termed BasicTAD), we decompose the TAD pipeline into several essential components: data sampling, backbone design, neck construction, and detection head. We extensively investigate the existing techniques in each component for this baseline, and more importantly, perform end-to-end training over the entire pipeline thanks to the simplicity of design. As a result, this simple BasicTAD yields an astounding and real-time RGB-Only baseline very close to the state-of-the-art methods with two-stream inputs. In addition, we further improve the BasicTAD by preserving more temporal and spatial information in network representation (termed as PlusTAD). Empirical results demonstrate that our PlusTAD is very efficient and significantly outperforms the previous methods on the datasets of THUMOS14 and FineAction. Meanwhile, we also perform in-depth visualization and error analysis on our proposed method and try to provide more insights on the TAD problem. Our approach can serve as a strong baseline for future TAD research. The code and model will be released at https://github.com/MCG-NJU/BasicTAD.

Guo Chen, Sen Xing, Zhe Chen et al.

In this report, we present our champion solutions to five tracks at Ego4D challenge. We leverage our developed InternVideo, a video foundation model, for five Ego4D tasks, including Moment Queries, Natural Language Queries, Future Hand Prediction, State Change Object Detection, and Short-term Object Interaction Anticipation. InternVideo-Ego4D is an effective paradigm to adapt the strong foundation model to the downstream ego-centric video understanding tasks with simple head designs. In these five tasks, the performance of InternVideo-Ego4D comprehensively surpasses the baseline methods and the champions of CVPR2022, demonstrating the powerful representation ability of InternVideo as a video foundation model. Our code will be released at https://github.com/OpenGVLab/ego4d-eccv2022-solutions

Shengyi Gao, Zhe Chen, Guo Chen et al.

The combination of audio and vision has long been a topic of interest in the multi-modal community. Recently, a new audio-visual segmentation (AVS) task has been introduced, aiming to locate and segment the sounding objects in a given video. This task demands audio-driven pixel-level scene understanding for the first time, posing significant challenges. In this paper, we propose AVSegFormer, a novel framework for AVS tasks that leverages the transformer architecture. Specifically, we introduce audio queries and learnable queries into the transformer decoder, enabling the network to selectively attend to interested visual features. Besides, we present an audio-visual mixer, which can dynamically adjust visual features by amplifying relevant and suppressing irrelevant spatial channels. Additionally, we devise an intermediate mask loss to enhance the supervision of the decoder, encouraging the network to produce more accurate intermediate predictions. Extensive experiments demonstrate that AVSegFormer achieves state-of-the-art results on the AVS benchmark. The code is available at https://github.com/vvvb-github/AVSegFormer.

Yi Wang, Yinan He, Yizhuo Li et al.

This paper introduces InternVid, a large-scale video-centric multimodal dataset that enables learning powerful and transferable video-text representations for multimodal understanding and generation. The InternVid dataset contains over 7 million videos lasting nearly 760K hours, yielding 234M video clips accompanied by detailed descriptions of total 4.1B words. Our core contribution is to develop a scalable approach to autonomously build a high-quality video-text dataset with large language models (LLM), thereby showcasing its efficacy in learning video-language representation at scale. Specifically, we utilize a multi-scale approach to generate video-related descriptions. Furthermore, we introduce ViCLIP, a video-text representation learning model based on ViT-L. Learned on InternVid via contrastive learning, this model demonstrates leading zero-shot action recognition and competitive video retrieval performance. Beyond basic video understanding tasks like recognition and retrieval, our dataset and model have broad applications. They are particularly beneficial for generating interleaved video-text data for learning a video-centric dialogue system, advancing video-to-text and text-to-video generation research. These proposed resources provide a tool for researchers and practitioners interested in multimodal video understanding and generation.

Jiahao Wang, Guo Chen, Yifei Huang et al.

Most existing forecasting systems are memory-based methods, which attempt to mimic human forecasting ability by employing various memory mechanisms and have progressed in temporal modeling for memory dependency. Nevertheless, an obvious weakness of this paradigm is that it can only model limited historical dependence and can not transcend the past. In this paper, we rethink the temporal dependence of event evolution and propose a novel memory-anticipation-based paradigm to model an entire temporal structure, including the past, present, and future. Based on this idea, we present Memory-and-Anticipation Transformer (MAT), a memory-anticipation-based approach, to address the online action detection and anticipation tasks. In addition, owing to the inherent superiority of MAT, it can process online action detection and anticipation tasks in a unified manner. The proposed MAT model is tested on four challenging benchmarks TVSeries, THUMOS'14, HDD, and EPIC-Kitchens-100, for online action detection and anticipation tasks, and it significantly outperforms all existing methods. Code is available at https://github.com/Echo0125/Memory-and-Anticipation-Transformer.

Shaobo Wang, Guo Chen, Ziyue Wang et al.

With the rapid progress of large language models (LLMs), reliably evaluating the capabilities of pre-trained LLMs has become increasingly important. The challenge is that base pre-trained models are optimized for next-token prediction and often fail to follow instructions or produce well-formed answers under standard prompting and direct decoding. As a result, benchmark performance can conflate model capability with decoding-induced failures to produce task-oriented outputs, while exposing such behavior often relies on costly post-training. Recent decodingonly approaches attempt to reshape output distributions, but such methods can be inefficient and brittle across open-ended tasks. To address these limitations, we propose Energy-Based Decoding (EBD), a training-free, reward-guided framework for activating task-oriented behaviors from frozen pre-trained LLMs across both open-ended and objective tasks. EBD augments decoding with an external lightweight reward model, steering generations toward high-utility responses while anchoring them to the pre-trained model prior through a reward-tilted target distribution. We show that EBD shifts base-model outputs toward more instructionfollowing behavior, increasing behavioral similarity to post-trained counterparts and enabling a fairer inference-time evaluation of accessible pre-trained-model behavior. Empirically, EBD outperforms baselines across five models and six benchmarks, improving Qwen3-8B-Base on AlpacaEval2.0 from 8.8 to 44.5, reducing Mistral-7B Math500 latency by 18.9x relative to prior decoding work, and remaining robust to reward-model size.

Amala Sanjay Deshmukh, Kateryna Chumachenko, Tuomas Rintamaki et al. · amazon-science, nvidia

We introduce Nemotron 3 Nano Omni, the latest model in the Nemotron multimodal series and the first to natively support audio inputs alongside text, images, and video. Nemotron 3 Nano Omni delivers consistent accuracy improvements over its predecessor, Nemotron Nano V2 VL, across all modalities, enabled by advances in architecture, training data and recipes. In particular, Nemotron 3 delivers leading results in real-world document understanding, long audio-video comprehension, and agentic computer use. Built on the highly efficient Nemotron 3 Nano 30B-A3B backbone, Nemotron 3 Nano Omni further incorporates innovative multimodal token-reduction techniques to deliver substantially lower inference latency and higher throughput than other models of similar size. We are releasing model checkpoints in BF16, FP8, and FP4 formats, along with portions of the training data and codebase to facilitate further research and development.

Shengyi Gao, Zhe Chen, Guo Chen et al.

In this report, we present our champion solution to the WSDM2023 Toloka Visual Question Answering (VQA) Challenge. Different from the common VQA and visual grounding (VG) tasks, this challenge involves a more complex scenario, i.e. inferring and locating the object implicitly specified by the given interrogative question. For this task, we leverage ViT-Adapter, a pre-training-free adapter network, to adapt multi-modal pre-trained Uni-Perceiver for better cross-modal localization. Our method ranks first on the leaderboard, achieving 77.5 and 76.347 IoU on public and private test sets, respectively. It shows that ViT-Adapter is also an effective paradigm for adapting the unified perception model to vision-language downstream tasks. Code and models will be released at https://github.com/czczup/ViT-Adapter/tree/main/wsdm2023.

Zhe Chen, Jiannan Wu, Wenhai Wang et al.

The exponential growth of large language models (LLMs) has opened up numerous possibilities for multimodal AGI systems. However, the progress in vision and vision-language foundation models, which are also critical elements of multi-modal AGI, has not kept pace with LLMs. In this work, we design a large-scale vision-language foundation model (InternVL), which scales up the vision foundation model to 6 billion parameters and progressively aligns it with the LLM, using web-scale image-text data from various sources. This model can be broadly applied to and achieve state-of-the-art performance on 32 generic visual-linguistic benchmarks including visual perception tasks such as image-level or pixel-level recognition, vision-language tasks such as zero-shot image/video classification, zero-shot image/video-text retrieval, and link with LLMs to create multi-modal dialogue systems. It has powerful visual capabilities and can be a good alternative to the ViT-22B. We hope that our research could contribute to the development of multi-modal large models. Code and models are available at https://github.com/OpenGVLab/InternVL.

Lingfeng Tang, Daoping Zhang, Junjie Chen et al.

Host-GPU data movement has become a latency-critical bottleneck in LLM serving, surfacing in common paths such as model-weight movement and KV cache offload/fetch. Today, each host-GPU copy is effectively confined to the PCIe path of the target GPU, even though modern multi-GPU servers contain additional PCIe links on peer GPUs and high bandwidth GPU interconnects. This leaves substantial intra-server I/O capacity unused. To address this issue, we present Multipath Memory Access (MMA), a software-defined multipath memory access system for host--GPU data transfer. To the best of our knowledge, MMA is the first software-defined system to enable efficient multipath host--GPU data transfer within a single multi-GPU server. MMA expands a single host--GPU copy across available direct and relay paths without hardware, driver, or application changes. It preserves CUDA stream semantics with a dependency-preserving Dummy Task, coordinates distributed micro-transfer completion through a lightweight synchronization mechanism, and uses queue backpressure to route traffic without explicit link-state feedback. On an 8-GPU NVIDIA H20 server, MMA achieves 245 GB/s peak host-to-GPU bandwidth, a 4.62x improvement over native CUDA copies, and reduces TTFT for KV cache fetching by 1.14-2.38x and model wake-up/switching latency by 1.12-2.48x.

Shihao Wang, Shilong Liu, Yuanguo Kuang et al.

Vision-language models (VLMs) commonly formulate visual grounding and detection as a coordinate-token generation problem, serializing each 2D box into multiple 1D tokens that are learned and decoded largely independently. This token-by-token decoding mismatches the coupled structure of box geometry and creates a practical inference bottleneck due to strictly sequential generation. We introduce LocateAnything, a unified generative grounding and detection framework based on Parallel Box Decoding (PBD). By decoding geometric elements such as bounding boxes and points as atomic units in a single step, LocateAnything preserves intra-box geometric coherence and unlocks substantial parallelism. We show that PBD improves both decoding throughput and localization accuracy. We further develop a scalable data engine and curate LocateAnything-Data, a large-scale dataset with more than 138 million training samples, substantially increasing data diversity for high-precision localization. Extensive evaluations show that LocateAnything advances the speed-accuracy frontier, achieving significantly higher decoding throughput while improving high-IoU localization quality across diverse benchmarks. The results highlight the complementary benefits of Parallel Box Decoding and large-scale training data in enabling efficient and precise unified visual grounding and detection.

Yi Wang, Kunchang Li, Xinhao Li et al.

We introduce InternVideo2, a new family of video foundation models (ViFM) that achieve the state-of-the-art results in video recognition, video-text tasks, and video-centric dialogue. Our core design is a progressive training approach that unifies the masked video modeling, crossmodal contrastive learning, and next token prediction, scaling up the video encoder size to 6B parameters. At the data level, we prioritize spatiotemporal consistency by semantically segmenting videos and generating video-audio-speech captions. This improves the alignment between video and text. Through extensive experiments, we validate our designs and demonstrate superior performance on over 60 video and audio tasks. Notably, our model outperforms others on various video-related dialogue and long video understanding benchmarks, highlighting its ability to reason and comprehend longer contexts. Code and models are available at https://github.com/OpenGVLab/InternVideo/tree/main/InternVideo2/.

Yin-Dong Zheng, Guo Chen, Minglei Yuan et al.

Action detection is a challenging video understanding task, requiring modeling spatio-temporal and interaction relations. Current methods usually model actor-actor and actor-context relations separately, ignoring their complementarity and mutual support. To solve this problem, we propose a novel network called Multi-Relation Support Network (MRSN). In MRSN, Actor-Context Relation Encoder (ACRE) and Actor-Actor Relation Encoder (AARE) model the actor-context and actor-actor relation separately. Then Relation Support Encoder (RSE) computes the supports between the two relations and performs relation-level interactions. Finally, Relation Consensus Module (RCM) enhances two relations with the long-term relations from the Long-term Relation Bank (LRB) and yields a consensus. Our experiments demonstrate that modeling relations separately and performing relation-level interactions can achieve and outperformer state-of-the-art results on two challenging video datasets: AVA and UCF101-24.

Yifei Huang, Guo Chen, Jilan Xu et al.

Being able to map the activities of others into one's own point of view is one fundamental human skill even from a very early age. Taking a step toward understanding this human ability, we introduce EgoExoLearn, a large-scale dataset that emulates the human demonstration following process, in which individuals record egocentric videos as they execute tasks guided by demonstration videos. Focusing on the potential applications in daily assistance and professional support, EgoExoLearn contains egocentric and demonstration video data spanning 120 hours captured in daily life scenarios and specialized laboratories. Along with the videos we record high-quality gaze data and provide detailed multimodal annotations, formulating a playground for modeling the human ability to bridge asynchronous procedural actions from different viewpoints. To this end, we present benchmarks such as cross-view association, cross-view action planning, and cross-view referenced skill assessment, along with detailed analysis. We expect EgoExoLearn can serve as an important resource for bridging the actions across views, thus paving the way for creating AI agents capable of seamlessly learning by observing humans in the real world. Code and data can be found at: https://github.com/OpenGVLab/EgoExoLearn

Shaobo Wang, Xuan Ouyang, Tianyi Xu et al.

As high-quality public text approaches exhaustion, a phenomenon known as the Data Wall, pre-training is shifting from more tokens to better tokens. However, existing methods either rely on heuristic static filters that ignore training dynamics, or use dynamic yet optimizer-agnostic criteria based on raw gradients. We propose OPUS (Optimizer-induced Projected Utility Selection), a dynamic data selection framework that defines utility in the optimizer-induced update space. OPUS scores candidates by projecting their effective updates, shaped by modern optimizers, onto a target direction derived from a stable, in-distribution proxy. To ensure scalability, we employ Ghost technique with CountSketch for computational efficiency, and Boltzmann sampling for data diversity, incurring only 4.7\% additional compute overhead. OPUS achieves remarkable results across diverse corpora, quality tiers, optimizers, and model scales. In pre-training of GPT-2 Large/XL on FineWeb and FineWeb-Edu with 30B tokens, OPUS outperforms industrial-level baselines and even full 200B-token training. Moreover, when combined with industrial-level static filters, OPUS further improves pre-training efficiency, even with lower-quality data. Furthermore, in continued pre-training of Qwen3-8B-Base on SciencePedia, OPUS achieves superior performance using only 0.5B tokens compared to full training with 3B tokens, demonstrating significant data efficiency gains in specialized domains.

Yinan He, Guo Chen

In this report, we present the transferring pretrained video mask autoencoders(VideoMAE) to egocentric tasks for Ego4d Looking at me Challenge. VideoMAE is the data-efficient pretraining model for self-supervised video pre-training and can easily transfer to downstream tasks. We show that the representation transferred from VideoMAE has good Spatio-temporal modeling and the ability to capture small actions. We only need to use egocentric data to train 10 epochs based on VideoMAE which pretrained by the ordinary videos acquired from a third person's view, and we can get better results than the baseline on Ego4d Looking at me Challenge.

Guo Chen, Yicheng Liu, Yifei Huang et al.

Most existing video understanding benchmarks for multimodal large language models (MLLMs) focus only on short videos. The limited number of benchmarks for long video understanding often rely solely on multiple-choice questions (MCQs). However, because of the inherent limitation of MCQ-based evaluation and the increasing reasoning ability of MLLMs, models can give the current answer purely by combining short video understanding with elimination, without genuinely understanding the video content. To address this gap, we introduce CG-Bench, a novel benchmark designed for clue-grounded question answering in long videos. CG-Bench emphasizes the model's ability to retrieve relevant clues for questions, enhancing evaluation credibility. It features 1,219 manually curated videos categorized by a granular system with 14 primary categories, 171 secondary categories, and 638 tertiary categories, making it the largest benchmark for long video analysis. The benchmark includes 12,129 QA pairs in three major question types: perception, reasoning, and hallucination. Compensating the drawbacks of pure MCQ-based evaluation, we design two novel clue-based evaluation methods: clue-grounded white box and black box evaluations, to assess whether the model generates answers based on the correct understanding of the video. We evaluate multiple closed-source and open-source MLLMs on CG-Bench. Results indicate that current models significantly underperform in understanding long videos compared to short ones, and a significant gap exists between open-source and commercial models. We hope CG-Bench can advance the development of more trustworthy and capable MLLMs for long video understanding. All annotations and video data are released at https://cg-bench.github.io/leaderboard/.

Zhiqi Li, Guo Chen, Shilong Liu et al.

Recently, promising progress has been made by open-source vision-language models (VLMs) in bringing their capabilities closer to those of proprietary frontier models. However, most open-source models only publish their final model weights, leaving the critical details of data strategies and implementation largely opaque. In this work, we address VLM post-training from a data-centric perspective, showing the key role of data strategy in developing frontier VLMs. By studying and building our post-training data strategy from scratch, we share detailed insights into the development processes, aiming to benefit the development of competitive models for the open-source community. Our introduced data strategy, together with training recipes and model design, leads to a family of performant VLMs named Eagle2. Specifically, Eagle2-9B achieves state-of-the-art results across various multimodal benchmarks, matching certain competitive models with up to 70B parameters.

Yin-Dong Zheng, Guo Chen, Jiahao Wang et al.

Capturing the state changes of interacting objects is a key technology for understanding human-object interactions. This technical report describes our method using heterogeneous backbones for the Ego4D Object State Change Classification and PNR Temporal Localization Challenge. In the challenge, we used the heterogeneous video understanding backbones, namely CSN with 3D convolution as operator and VideoMAE with Transformer as operator. Our method achieves an accuracy of 0.796 on OSCC while achieving an absolute temporal localization error of 0.516 on PNR. These excellent results rank 1st on the leaderboard of Ego4D OSCC & PNR-TL Challenge 2022.

Fanglin Xu, Wei Zhang, Jian Yang et al.

The rapid advancement of code large language models (LLMs) has sparked significant research interest in systematically evaluating their code generation capabilities, yet existing benchmarks predominantly assess models at a single structural granularity and focus on limited programming languages, obscuring fine-grained capability variations across different code scopes and multilingual scenarios. We introduce M2G-Eval, a multi-granularity, multilingual framework for evaluating code generation in large language models (LLMs) across four levels: Class, Function, Block, and Line. Spanning 18 programming languages, M2G-Eval includes 17K+ training tasks and 1,286 human-annotated, contamination-controlled test instances. We develop M2G-Eval-Coder models by training Qwen3-8B with supervised fine-tuning and Group Relative Policy Optimization. Evaluating 30 models (28 state-of-the-art LLMs plus our two M2G-Eval-Coder variants) reveals three main findings: (1) an apparent difficulty hierarchy, with Line-level tasks easiest and Class-level most challenging; (2) widening performance gaps between full- and partial-granularity languages as task complexity increases; and (3) strong cross-language correlations, suggesting that models learn transferable programming concepts. M2G-Eval enables fine-grained diagnosis of code generation capabilities and highlights persistent challenges in synthesizing complex, long-form code.

Guo Chen, Zhiqi Li, Shihao Wang et al.

We introduce Eagle 2.5, a family of frontier vision-language models (VLMs) for long-context multimodal learning. Our work addresses the challenges in long video comprehension and high-resolution image understanding, introducing a generalist framework for both tasks. The proposed training framework incorporates Automatic Degrade Sampling and Image Area Preservation, two techniques that preserve contextual integrity and visual details. The framework also includes numerous efficiency optimizations in the pipeline for long-context data training. Finally, we propose Eagle-Video-110K, a novel dataset that integrates both story-level and clip-level annotations, facilitating long-video understanding. Eagle 2.5 demonstrates substantial improvements on long-context multimodal benchmarks, providing a robust solution to the limitations of existing VLMs. Notably, our best model Eagle 2.5-8B achieves 72.4% on Video-MME with 512 input frames, matching the results of top-tier commercial model such as GPT-4o and large-scale open-source models like Qwen2.5-VL-72B and InternVL2.5-78B.

Yifei Huang, Jilan Xu, Baoqi Pei et al.

We introduce Vinci, a real-time embodied smart assistant built upon an egocentric vision-language model. Designed for deployment on portable devices such as smartphones and wearable cameras, Vinci operates in an "always on" mode, continuously observing the environment to deliver seamless interaction and assistance. Users can wake up the system and engage in natural conversations to ask questions or seek assistance, with responses delivered through audio for hands-free convenience. With its ability to process long video streams in real-time, Vinci can answer user queries about current observations and historical context while also providing task planning based on past interactions. To further enhance usability, Vinci integrates a video generation module that creates step-by-step visual demonstrations for tasks that require detailed guidance. We hope that Vinci can establish a robust framework for portable, real-time egocentric AI systems, empowering users with contextual and actionable insights. We release the complete implementation for the development of the device in conjunction with a demo web platform to test uploaded videos at https://github.com/OpenGVLab/vinci.

Yi Li, Han Liu, Mingfeng Fan et al.

Federated learning (FL) on graphs shows promise for distributed time-series forecasting. Yet, existing methods rely on static topologies and struggle with client heterogeneity. We propose Fed-GAME, a framework that models personalized aggregation as message passing over a learnable dynamic implicit graph. The core is a decoupled parameter difference-based update protocol, where clients transmit parameter differences between their fine-tuned private model and a shared global model. On the server, these differences are decomposed into two streams: (1) averaged difference used to updating the global model for consensus (2) the selective difference fed into a novel Graph Attention Mixture-of-Experts (GAME) aggregator for fine-grained personalization. In this aggregator, shared experts provide scoring signals while personalized gates adaptively weight selective updates to support personalized aggregation. Experiments on two real-world electric vehicle charging datasets demonstrate that Fed-GAME outperforms state-of-the-art personalized FL baselines.

Baoqi Pei, Yifei Huang, Jilan Xu et al.

In egocentric video understanding, the motion of hands and objects as well as their interactions play a significant role by nature. However, existing egocentric video representation learning methods mainly focus on aligning video representation with high-level narrations, overlooking the intricate dynamics between hands and objects. In this work, we aim to integrate the modeling of fine-grained hand-object dynamics into the video representation learning process. Since no suitable data is available, we introduce HOD, a novel pipeline employing a hand-object detector and a large language model to generate high-quality narrations with detailed descriptions of hand-object dynamics. To learn these fine-grained dynamics, we propose EgoVideo, a model with a new lightweight motion adapter to capture fine-grained hand-object motion information. Through our co-training strategy, EgoVideo effectively and efficiently leverages the fine-grained hand-object dynamics in the HOD data. Extensive experiments demonstrate that our method achieves state-of-the-art performance across multiple egocentric downstream tasks, including improvements of 6.3% in EK-100 multi-instance retrieval, 5.7% in EK-100 classification, and 16.3% in EGTEA classification in zero-shot settings. Furthermore, our model exhibits robust generalization capabilities in hand-object interaction and robot manipulation tasks. Code and data are available at https://github.com/OpenRobotLab/EgoHOD/.

Xi Zhang, Huqiang Cheng, Guo Chen et al.

Peer-to-peer (P2P) energy management facilitates decentralized resource allocation among prosumers, improving local hosting capacity for renewables and minimizing energy expenditures while ensuring data privacy through distributed coordination. However, conventional P2P energy management methods are confined to synchronous scheduling paradigms, creating synchronization bottlenecks that fundamentally conflict with the dynamic and decentralized nature of P2P energy management tasks. To bridge this gap, this paper focuses on resolving a class of dynamic energy management problems over asynchronous P2P (Asyn-P2P) transactive networks. We first recast the dynamic energy management problems into a saddle-point problem, and then propose a synchronous decentralized dynamic energy management algorithm, dubbed Syn-DYNA,based on operator splitting theory. To eliminate the global synchronization clock in Syn-DYNA, we introduce a random activation scheme, together with local buffers for latest state tracking, to develop an asynchronous variant of Syn-DYNA, namely Asyn-DYNA. Based on monotone operator theory, theoretical analysis proves a non-asymptotic linear convergence rate for Syn-DYNA and establishes the almost sure convergence ofAsyn-DYNA. Numerical experiments validate effectiveness of Syn-DYNA and Asyn-DYNA algorithms by tackling a dynamic energy management task over P2P transactive networks.

Shaobo Wang, Yantai Yang, Guo Chen et al.

Dataset Distillation (DD) seeks to create a compact dataset from a large, real-world dataset. While recent methods often rely on heuristic approaches to balance efficiency and quality, the fundamental relationship between original and synthetic data remains underexplored. This paper revisits knowledge distillation-based dataset distillation within a solid theoretical framework. We introduce the concepts of Informativeness and Utility, capturing crucial information within a sample and essential samples in the training set, respectively. Building on these principles, we define optimal dataset distillation mathematically. We then present InfoUtil, a framework that balances informativeness and utility in synthesizing the distilled dataset. InfoUtil incorporates two key components: (1) game-theoretic informativeness maximization using Shapley Value attribution to extract key information from samples, and (2) principled utility maximization by selecting globally influential samples based on Gradient Norm. These components ensure that the distilled dataset is both informative and utility-optimized. Experiments demonstrate that our method achieves a 6.1\% performance improvement over the previous state-of-the-art approach on ImageNet-1K dataset using ResNet-18.

Yifei Huang, Jilan Xu, Baoqi Pei et al.

We present Vinci, a vision-language system designed to provide real-time, comprehensive AI assistance on portable devices. At its core, Vinci leverages EgoVideo-VL, a novel model that integrates an egocentric vision foundation model with a large language model (LLM), enabling advanced functionalities such as scene understanding, temporal grounding, video summarization, and future planning. To enhance its utility, Vinci incorporates a memory module for processing long video streams in real time while retaining contextual history, a generation module for producing visual action demonstrations, and a retrieval module that bridges egocentric and third-person perspectives to provide relevant how-to videos for skill acquisition. Unlike existing systems that often depend on specialized hardware, Vinci is hardware-agnostic, supporting deployment across a wide range of devices, including smartphones and wearable cameras. In our experiments, we first demonstrate the superior performance of EgoVideo-VL on multiple public benchmarks, showcasing its vision-language reasoning and contextual understanding capabilities. We then conduct a series of user studies to evaluate the real-world effectiveness of Vinci, highlighting its adaptability and usability in diverse scenarios. We hope Vinci can establish a new framework for portable, real-time egocentric AI systems, empowering users with contextual and actionable insights. Including the frontend, backend, and models, all codes of Vinci are available at https://github.com/OpenGVLab/vinci.

Hongxiang Jiang, Jihao Yin, Qixiong Wang et al.

Recent advances in multimodal large language models (MLLMs) have demonstrated impressive results in various visual tasks. However, in remote sensing (RS), high resolution and small proportion of objects pose challenges to existing MLLMs, which struggle with object-centric tasks, particularly in precise localization and fine-grained attribute description for each object. These RS MLLMs have not yet surpassed classical visual perception models, as they only provide coarse image understanding, leading to limited gains in real-world scenarios. To address this gap, we establish EagleVision, an MLLM tailored for remote sensing that excels in object detection and attribute comprehension. Equipped with the Attribute Disentangle module, EagleVision learns disentanglement vision tokens to express distinct attributes. To support object-level visual-language alignment, we construct EVAttrs-95K, the first large-scale object attribute understanding dataset in RS for instruction tuning, along with a novel evaluation benchmark, EVBench. EagleVision achieves state-of-the-art performance on both fine-grained object detection and object attribute understanding tasks, highlighting the mutual promotion between detection and understanding capabilities in MLLMs. The code, model, data, and demo will be available at https://github.com/XiangTodayEatsWhat/EagleVision.

Yipeng Zou, Qin Liu, Jie Wu et al.

Ensemble attacks integrate the outputs of surrogate models with diverse architectures, which can be combined with various gradient-based attacks to improve adversarial transferability. However, previous work shows unsatisfactory attack performance when transferring across heterogeneous model architectures. The main reason is that the gradient update directions of heterogeneous surrogate models differ widely, making it hard to reduce the gradient variance of ensemble models while making the best of individual model. To tackle this challenge, we design a novel ensemble attack, NAMEA, which for the first time integrates the gradients from the non-attention areas of ensemble models into the iterative gradient optimization process. Our design is inspired by the observation that the attention areas of heterogeneous models vary sharply, thus the non-attention areas of ViTs are likely to be the focus of CNNs and vice versa. Therefore, we merge the gradients respectively from the attention and non-attention areas of ensemble models so as to fuse the transfer information of CNNs and ViTs. Specifically, we pioneer a new way of decoupling the gradients of non-attention areas from those of attention areas, while merging gradients by meta-learning. Empirical evaluations on ImageNet dataset indicate that NAMEA outperforms AdaEA and SMER, the state-of-the-art ensemble attacks by an average of 15.0% and 9.6%, respectively. This work is the first attempt to explore the power of ensemble non-attention in boosting cross-architecture transferability, providing new insights into launching ensemble attacks.

Lidong Lu, Guo Chen, Zhu Wei et al.

We introduce Audio-Visual Affordance Grounding (AV-AG), a new task that segments object interaction regions from action sounds. Unlike existing approaches that rely on textual instructions or demonstration videos, which often limited by ambiguity or occlusion, audio provides real-time, semantically rich, and visually independent cues for affordance grounding, enabling more intuitive understanding of interaction regions. To support this task, we construct the first AV-AG dataset, comprising a large collection of action sounds, object images, and pixel-level affordance annotations. The dataset also includes an unseen subset to evaluate zero-shot generalization. Furthermore, we propose AVAGFormer, a model equipped with a semantic-conditioned cross-modal mixer and a dual-head decoder that effectively fuses audio and visual signals for mask prediction. Experiments show that AVAGFormer achieves state-of-the-art performance on AV-AG, surpassing baselines from related tasks. Comprehensive analyses highlight the distinctions between AV-AG and AVS, the benefits of end-to-end modeling, and the contribution of each component. Code and dataset have been released on https://jscslld.github.io/AVAGFormer/.

Xu Yang, Jiapeng Zhang, Dongyang Zhao et al.

The KV cache in self-attention has emerged as a major bottleneck in long-context and large-batch inference for LLMs. Existing approaches often treat sparsity prediction and compression as separate modules, relying on auxiliary index structures to select relevant tokens, and on complex quantization schemes to reduce memory usage. This fragmented design introduces redundant overhead and limits scalability. In this paper, we propose a novel paradigm: treating the compressed key representation not merely as storage, but as a self-indexing structure that directly enables efficient sparse attention. By designing a sign-based 1-bit vector quantization (VQ) scheme, our method unifies compression and retrieval in a single, hardware-friendly format. This approach eliminates the need for external indices or learning-based predictors, offering a lightweight yet robust solution for memory-constrained inference. All components are designed to be hardware-efficient and easy to implement. By implementing custom CUDA kernels, our method integrates seamlessly with FlashAttention, minimizing additional runtime and memory overhead. Experimental results demonstrate that our approach delivers both effectiveness and efficiency.

Baoqi Pei, Yifei Huang, Jilan Xu et al.

Egocentric video reasoning centers on an unobservable agent behind the camera who dynamically shapes the environment, requiring inference of hidden intentions and recognition of fine-grained interactions. This core challenge limits current multimodal large language models MLLMs, which excel at visible event reasoning but lack embodied, first-person understanding. To bridge this gap, we introduce EgoThinker, a novel framework that endows MLLMs with robust egocentric reasoning capabilities through spatio-temporal chain-of-thought supervision and a two-stage learning curriculum. First, we introduce EgoRe-5M, a large-scale egocentric QA dataset constructed from 13M diverse egocentric video clips. This dataset features multi-minute segments annotated with detailed CoT rationales and dense hand-object grounding. Second, we employ SFT on EgoRe-5M to instill reasoning skills, followed by reinforcement fine-tuning RFT to further enhance spatio-temporal localization. Experimental results show that EgoThinker outperforms existing methods across multiple egocentric benchmarks, while achieving substantial improvements in fine-grained spatio-temporal localization tasks. Full code and data are released at https://github.com/InternRobotics/EgoThinker.

Qin Wang, Youhuan Li, Yansong Feng et al.

People without a database background usually rely on file systems or tools such as Excel for data management, which often lead to redundancy and data inconsistency. Relational databases possess strong data management capabilities, but require a high level of professional expertise from users. Although there are already many works on Text2SQL to automate the translation of natural language into SQL queries for data manipulation, all of them presuppose that the database schema is pre-designed. In practice, schema design itself demands domain expertise, and research on directly generating schemas from textual requirements remains unexplored. In this paper, we systematically define a new problem, called Text2Schema, to convert a natural language text requirement into a relational database schema. With an effective Text2Schema technique, users can effortlessly create database table structures using natural language, and subsequently leverage existing Text2SQL techniques to perform data manipulations, which significantly narrows the gap between non-technical personnel and highly efficient, versatile relational database systems. We propose SchemaAgent, an LLM-based multi-agent framework for Text2Schema. We emulate the workflow of manual schema design by assigning specialized roles to agents and enabling effective collaboration to refine their respective subtasks. We also incorporate dedicated roles for reflection and inspection, along with an innovative error detection and correction mechanism to identify and rectify issues across various phases. Moreover, we build and open source a benchmark containing 381 pairs of requirement description and schema. Experimental results demonstrate the superiority of our approach over comparative work.

Baoqi Pei, Guo Chen, Jilan Xu et al.

In this report, we present our solutions to the EgoVis Challenges in CVPR 2024, including five tracks in the Ego4D challenge and three tracks in the EPIC-Kitchens challenge. Building upon the video-language two-tower model and leveraging our meticulously organized egocentric video data, we introduce a novel foundation model called EgoVideo. This model is specifically designed to cater to the unique characteristics of egocentric videos and provides strong support for our competition submissions. In the Ego4D challenges, we tackle various tasks including Natural Language Queries, Step Grounding, Moment Queries, Short-term Object Interaction Anticipation, and Long-term Action Anticipation. In addition, we also participate in the EPIC-Kitchens challenge, where we engage in the Action Recognition, Multiple Instance Retrieval, and Domain Adaptation for Action Recognition tracks. By adapting EgoVideo to these diverse tasks, we showcase its versatility and effectiveness in different egocentric video analysis scenarios, demonstrating the powerful representation ability of EgoVideo as an egocentric foundation model. Our codebase and pretrained models are publicly available at https://github.com/OpenGVLab/EgoVideo.

Guo Chen, Yifei Huang, Jilan Xu et al.

Understanding videos is one of the fundamental directions in computer vision research, with extensive efforts dedicated to exploring various architectures such as RNN, 3D CNN, and Transformers. The newly proposed architecture of state space model, e.g., Mamba, shows promising traits to extend its success in long sequence modeling to video modeling. To assess whether Mamba can be a viable alternative to Transformers in the video understanding domain, in this work, we conduct a comprehensive set of studies, probing different roles Mamba can play in modeling videos, while investigating diverse tasks where Mamba could exhibit superiority. We categorize Mamba into four roles for modeling videos, deriving a Video Mamba Suite composed of 14 models/modules, and evaluating them on 12 video understanding tasks. Our extensive experiments reveal the strong potential of Mamba on both video-only and video-language tasks while showing promising efficiency-performance trade-offs. We hope this work could provide valuable data points and insights for future research on video understanding. Code is public: https://github.com/OpenGVLab/video-mamba-suite.

Guo Chen, Yin-Dong Zheng, Jiahao Wang et al.

With the exponential growth of video data, there is an urgent need for automated technology to analyze and comprehend video content. However, existing video understanding models are often task-specific and lack a comprehensive capability of handling diverse tasks. The success of large language models (LLMs) like GPT has demonstrated their impressive abilities in sequence causal reasoning. Building upon this insight, we propose a novel framework called VideoLLM that leverages the sequence reasoning capabilities of pre-trained LLMs from natural language processing (NLP) for video sequence understanding. VideoLLM incorporates a carefully designed Modality Encoder and Semantic Translator, which convert inputs from various modalities into a unified token sequence. This token sequence is then fed into a decoder-only LLM. Subsequently, with the aid of a simple task head, our VideoLLM yields an effective unified framework for different kinds of video understanding tasks. To evaluate the efficacy of VideoLLM, we conduct extensive experiments using multiple LLMs and fine-tuning methods. We evaluate our VideoLLM on eight tasks sourced from four different datasets. The experimental results demonstrate that the understanding and reasoning capabilities of LLMs can be effectively transferred to video understanding tasks. We release the code at https://github.com/cg1177/VideoLLM.

Guo Chen, Yin-Dong Zheng, Limin Wang et al.

Temporal action detection aims to locate the boundaries of action in the video. The current method based on boundary matching enumerates and calculates all possible boundary matchings to generate proposals. However, these methods neglect the long-range context aggregation in boundary prediction. At the same time, due to the similar semantics of adjacent matchings, local semantic aggregation of densely-generated matchings cannot improve semantic richness and discrimination. In this paper, we propose the end-to-end proposal generation method named Dual Context Aggregation Network (DCAN) to aggregate context on two levels, namely, boundary level and proposal level, for generating high-quality action proposals, thereby improving the performance of temporal action detection. Specifically, we design the Multi-Path Temporal Context Aggregation (MTCA) to achieve smooth context aggregation on boundary level and precise evaluation of boundaries. For matching evaluation, Coarse-to-fine Matching (CFM) is designed to aggregate context on the proposal level and refine the matching map from coarse to fine. We conduct extensive experiments on ActivityNet v1.3 and THUMOS-14. DCAN obtains an average mAP of 35.39% on ActivityNet v1.3 and reaches mAP 54.14% at IoU@0.5 on THUMOS-14, which demonstrates DCAN can generate high-quality proposals and achieve state-of-the-art performance. We release the code at https://github.com/cg1177/DCAN.

Zhe Chen, Jiahao Wang, Wenhai Wang et al.

We propose an accurate and efficient scene text detection framework, termed FAST (i.e., faster arbitrarily-shaped text detector). Different from recent advanced text detectors that used complicated post-processing and hand-crafted network architectures, resulting in low inference speed, FAST has two new designs. (1) We design a minimalist kernel representation (only has 1-channel output) to model text with arbitrary shape, as well as a GPU-parallel post-processing to efficiently assemble text lines with a negligible time overhead. (2) We search the network architecture tailored for text detection, leading to more powerful features than most networks that are searched for image classification. Benefiting from these two designs, FAST achieves an excellent trade-off between accuracy and efficiency on several challenging datasets, including Total Text, CTW1500, ICDAR 2015, and MSRA-TD500. For example, FAST-T yields 81.6% F-measure at 152 FPS on Total-Text, outperforming the previous fastest method by 1.7 points and 70 FPS in terms of accuracy and speed. With TensorRT optimization, the inference speed can be further accelerated to over 600 FPS. Code and models will be released at https://github.com/czczup/FAST.

Jilan Xu, Yifei Huang, Junlin Hou et al.

Understanding human actions from videos of first-person view poses significant challenges. Most prior approaches explore representation learning on egocentric videos only, while overlooking the potential benefit of exploiting existing large-scale third-person videos. In this paper, (1) we develop EgoInstructor, a retrieval-augmented multimodal captioning model that automatically retrieves semantically relevant third-person instructional videos to enhance the video captioning of egocentric videos. (2) For training the cross-view retrieval module, we devise an automatic pipeline to discover ego-exo video pairs from distinct large-scale egocentric and exocentric datasets. (3) We train the cross-view retrieval module with a novel EgoExoNCE loss that pulls egocentric and exocentric video features closer by aligning them to shared text features that describe similar actions. (4) Through extensive experiments, our cross-view retrieval module demonstrates superior performance across seven benchmarks. Regarding egocentric video captioning, EgoInstructor exhibits significant improvements by leveraging third-person videos as references. Project page is available at: https://jazzcharles.github.io/Egoinstructor/

Jindong Jiang, Xiuyu Li, Zhijian Liu et al.

Recent advances in video-based multimodal large language models (Video-LLMs) have significantly improved video understanding by processing videos as sequences of image frames. However, many existing methods treat frames independently in the vision backbone, lacking explicit temporal modeling, which limits their ability to capture dynamic patterns and efficiently handle long videos. To address these limitations, we introduce STORM (Spatiotemporal TOken Reduction for Multimodal LLMs), a novel architecture incorporating a dedicated temporal encoder between the image encoder and the LLM. Our temporal encoder leverages the Mamba State Space Model to integrate temporal information into image tokens, generating enriched representations that preserve inter-frame dynamics across the entire video sequence. This enriched encoding not only enhances video reasoning capabilities but also enables effective token reduction strategies, including test-time sampling and training-based temporal and spatial pooling, substantially reducing computational demands on the LLM without sacrificing key temporal information. By integrating these techniques, our approach simultaneously reduces training and inference latency while improving performance, enabling efficient and robust video understanding over extended temporal contexts. Extensive evaluations show that STORM achieves state-of-the-art results across various long video understanding benchmarks (more than 5% improvement on MLVU and LongVideoBench) while reducing the computation costs by up to $8\times$ and the decoding latency by 2.4-2.9$\times$ for the fixed numbers of input frames. Project page is available at https://research.nvidia.com/labs/lpr/storm

Jilan Xu, Yifei Huang, Baoqi Pei et al.

Generating videos in the first-person perspective has broad application prospects in the field of augmented reality and embodied intelligence. In this work, we explore the cross-view video prediction task, where given an exo-centric video, the first frame of the corresponding ego-centric video, and textual instructions, the goal is to generate futur frames of the ego-centric video. Inspired by the notion that hand-object interactions (HOI) in ego-centric videos represent the primary intentions and actions of the current actor, we present EgoExo-Gen that explicitly models the hand-object dynamics for cross-view video prediction. EgoExo-Gen consists of two stages. First, we design a cross-view HOI mask prediction model that anticipates the HOI masks in future ego-frames by modeling the spatio-temporal ego-exo correspondence. Next, we employ a video diffusion model to predict future ego-frames using the first ego-frame and textual instructions, while incorporating the HOI masks as structural guidance to enhance prediction quality. To facilitate training, we develop an automated pipeline to generate pseudo HOI masks for both ego- and exo-videos by exploiting vision foundation models. Extensive experiments demonstrate that our proposed EgoExo-Gen achieves better prediction performance compared to previous video prediction models on the Ego-Exo4D and H2O benchmark datasets, with the HOI masks significantly improving the generation of hands and interactive objects in the ego-centric videos.

Guo Chen, Lidong Lu, Yicheng Liu et al.

While datasets for video understanding have scaled to hour-long durations, they typically consist of densely concatenated clips that differ from natural, unscripted daily life. To bridge this gap, we introduce MM-Lifelong, a dataset designed for Multimodal Lifelong Understanding. Comprising 181.1 hours of footage, it is structured across Day, Week, and Month scales to capture varying temporal densities. Extensive evaluations reveal two critical failure modes in current paradigms: end-to-end MLLMs suffer from a Working Memory Bottleneck due to context saturation, while representative agentic baselines experience Global Localization Collapse when navigating sparse, month-long timelines. To address this, we propose the Recursive Multimodal Agent (ReMA), which employs dynamic memory management to iteratively update a recursive belief state, significantly outperforming existing methods. Finally, we establish dataset splits designed to isolate temporal and domain biases, providing a rigorous foundation for future research in supervised learning and out-of-distribution generalization.

Zihang Qiu, Chaojie Li, Zhongyang Wang et al.

Accurate prediction helps to achieve supply-demand balance in energy systems, supporting decision-making and scheduling. Traditional models, lacking AI-assisted automation, rely on experts, incur high costs, and struggle with sparse data prediction. To address these challenges, we propose the Energy Forecasting Large Language Model (EF-LLM), which integrates domain knowledge and temporal data for time-series forecasting, supporting both pre-forecast operations and post-forecast decision-support. EF-LLM's human-AI interaction capabilities lower the entry barrier in forecasting tasks, reducing the need for extra expert involvement. To achieve this, we propose a continual learning approach with updatable LoRA and a multi-channel architecture for aligning heterogeneous multimodal data, enabling EF-LLM to continually learn heterogeneous multimodal knowledge. In addition, EF-LLM enables accurate predictions under sparse data conditions through its ability to process multimodal data. We propose Fusion Parameter-Efficient Fine-Tuning (F-PEFT) method to effectively leverage both time-series data and text for this purpose. EF-LLM is also the first energy-specific LLM to detect hallucinations and quantify their occurrence rate, achieved via multi-task learning, semantic similarity analysis, and ANOVA. We have achieved success in energy prediction scenarios for load, photovoltaic, and wind power forecast.

Shihao Wang, Guo Chen, De-an Huang et al.

Recent studies have revealed that selecting informative and relevant video frames can significantly improve the performance of Video Large Language Models (Video-LLMs). Current methods, such as reducing inter-frame redundancy, employing separate models for image-text relevance assessment, or utilizing temporal video grounding for event localization, substantially adopt unsupervised learning paradigms, whereas they struggle to address the complex scenarios in long video understanding. We propose Instructed Temporal Grounding for Videos (VideoITG), featuring customized frame sampling aligned with user instructions. The core of VideoITG is the VidThinker pipeline, an automated annotation framework that explicitly mimics the human annotation process. First, it generates detailed clip-level captions conditioned on the instruction; then, it retrieves relevant video segments through instruction-guided reasoning; finally, it performs fine-grained frame selection to pinpoint the most informative visual evidence. Leveraging VidThinker, we construct the VideoITG-40K dataset, containing 40K videos and 500K instructed temporal grounding annotations. We then design a plug-and-play VideoITG model, which takes advantage of visual language alignment and reasoning capabilities of Video-LLMs, for effective frame selection in a discriminative manner. Coupled with Video-LLMs, VideoITG achieves consistent performance improvements across multiple multimodal video understanding benchmarks, showing its superiority and great potentials for video understanding.

Lidong Lu, Guo Chen, Zhiqi Li et al.

Despite progress in video understanding, current MLLMs struggle with counting tasks. Existing benchmarks are limited by short videos, close-set queries, lack of clue annotations, and weak multimodal coverage. In this paper, we introduce CG-AV-Counting, a manually-annotated clue-grounded counting benchmark with 1,027 multimodal questions and 5,845 annotated clues over 497 long videos. It supports both black-box and white-box evaluation, serving as a comprehensive testbed for both end-to-end and reasoning-based counting. To explore ways to improve model's counting capability, we propose AV-Reasoner, a model trained with GRPO and curriculum learning to generalize counting ability from related tasks. AV-Reasoner achieves state-of-the-art results across multiple benchmarks, demonstrating the effectiveness of reinforcement learning. However, experiments show that on out-of-domain benchmarks, reasoning in the language space fails to bring performance gains. The code and benchmark have been released on https://av-reasoner.github.io.

Yuping He, Yifei Huang, Guo Chen et al.

Transferring and integrating knowledge across first-person (egocentric) and third-person (exocentric) viewpoints is intrinsic to human intelligence, enabling humans to learn from others and convey insights from their own experiences. Despite rapid progress in multimodal large language models (MLLMs), their ability to perform such cross-view reasoning remains unexplored. To address this, we introduce EgoExoBench, the first benchmark for egocentric-exocentric video understanding and reasoning. Built from publicly available datasets, EgoExoBench comprises over 7,300 question-answer pairs spanning eleven sub-tasks organized into three core challenges: semantic alignment, viewpoint association, and temporal reasoning. We evaluate 13 state-of-the-art MLLMs and find that while these models excel on single-view tasks, they struggle to align semantics across perspectives, accurately associate views, and infer temporal dynamics in the ego-exo context. We hope EgoExoBench can serve as a valuable resource for research on embodied agents and intelligent assistants seeking human-like cross-view intelligence.