搜索:VLM

共命中 50 条(服务端检索)
多模态推理的基础设施:VL 模型服务化的特殊工程
视觉语言模型(VLM)的服务化与纯文本模型不同:图像的预处理管线、多模态 KV 的显存压力、请求的异构性(图+文的组合)。本篇拆解多模态推理基础设施的特殊工程点——从请求处理到显存管理到成本核算。
大模型 精选 · 原创 · 11-4 阅读 53·访客 53
In-Context Robot Learning with VLM Agents
Enabling robots to adapt to unfamiliar environments as readily as humans remains a moonshot goal of embodied AI. No fini…
智能体 HuggingFace Daily Papers · 9-16 阅读 90·访客 87
Show-Harness: Just a VLM Agent Can Play Robots
Foundation vision-language models (VLMs) exhibit broad intelligence about the world, yet translating this intelligence i…
智能体 HuggingFace Daily Papers · 9-9 阅读 100·访客 98
视频理解研究:多模态视频理解的研究进展与基准
从 Kinetics 到 Video-MME 的基准演进、视频编码器到 VLM 的模型迭代、token 瓶颈与帧采样难题:多模态视频理解的进展全景与开放问题。
研究前沿 精选 · 原创 · 11-9 阅读 57·访客 57
一篇读懂文档解析:RAG 的质量,卡在切块之前
RAG 答案不对,先别急着调切块参数——病灶常在更上游的 PDF 解析。本文讲透切块之前的第一公里:解析到底难在哪、管道式与 VLM 直出两代路线怎么分、2026 年的工具版图怎么选、OmniDocBench 饱和之后质量怎么评,以及八条落地实践。
一叶一世界 精选 · 原创 · 10-24 阅读 186·访客 177
Robo-COP:机器人编排器与策略在部署中共进化
论文提出VLA策略在训练域外泛化能力有限、成为系统瓶颈,提出Robo-COP方法让VLM编排器与策略在部署中协同进化,以克服单独微调策略导致与编排器脱节的问题。
研究前沿 HuggingFace Daily Papers · 10-6 阅读 39·访客 39
一篇读懂视觉语言模型:图像是怎么变成「语言」的
大语言模型只认 token 序列,图片如何进入对话?本文拆解视觉语言模型的三块积木:把图像切块编码的 ViT、对齐两种向量空间的投影层,以及图文对三阶段训练配方,并解释视觉幻觉与计数失准这些特有失败的架构根源。
一叶一世界 精选 · 原创 · 10-6 阅读 577·访客 571
Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success Rate but 65% Fewer Tokens
Vision language model (VLM) agents can control robots through visual feedback and action primitives, but repeated model …
智能体 HuggingFace Daily Papers · 10-1 阅读 70·访客 69
AnswerMap: Faithful Spatial Interpretability of VLMs from Answer Posteriors
When a VLM answers a visual query, current interpretability tools rely on text rationales, which use a mismatched modali…
行业动态 HuggingFace Daily Papers · 9-28 阅读 49·访客 48
CameraVLM:让视觉语言模型理解相机运动方向
论文指出多数VLM在判断相机左移还是右移时接近随机猜测,并提出CameraVLM,通过相机运动令牌与屏幕内光流令牌的联合表征以及状态-语言对齐来提升相机运动理解能力。
研究前沿 HuggingFace Daily Papers · 10-12 阅读 32·访客 32
PaperOrchestra:Google 五步多智能体流水线,把实验日志写成可投稿的 LaTeX 论文
Google Cloud AI Research 开源的 PaperOrchestra 用五个专职智能体把零散的想法摘要与实验日志,写成带真实文献引用、自动绘图、符合 CVPR 与 ICLR 模板的投稿级全稿,单篇约 40 分钟。本文拆解五步流水线架构、Semantic Scholar 引用认证机制与 PaperWritingBench 实测数据。
开源项目 精选 · 用户投稿 · 2天前 阅读 7·访客 7
ScaleVLA: Scaling Vision-Language-Action Models to 552 Billion Parameters
Frontier Vision-Language Models (VLMs) such as GPT-6 Astra have demonstrated remarkable capabilities in robot control, s…
智能体 HuggingFace Daily Papers · 5天前 阅读 2·访客 2
V-JEPA-LM: Video-Language Models That Read Dynamics Beyond Static, Text-Aligned Representations
World-model video encoders such as V-JEPA learn how scenes change (direction, speed, order, state) from video alone. To …
行业动态 HuggingFace Daily Papers · 11-13 阅读 2·访客 2
VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender
Controllable real-world video generation models can condition on a rendered graybox proxy video, which fixes layout, mot…
智能体 HuggingFace Daily Papers · 11-11 阅读 4·访客 4
Ego Guidance: Enhancing Robot Policy Learning with Action-Free Human Egocentric Videos
In this paper, we investigate how human egocentric videos can benefit robot policy learning without action supervision. …
智能体 HuggingFace Daily Papers · 11-10 阅读 3·访客 3
苹果罕见公布人形机器人进展,宇树 G1 背着一台跑千问的 Mac mini
苹果新 CEO 最近对自家 AI 的进度有点不满意。 据彭博社最新爆料,苹果 CEO 约翰·特努斯已经要求软件主管费德里吉加快 AI 开发节奏,拿出更大胆的方案,不能只满足于给 Siri 升级几个功能。 比起苹果 AI 落地的速度,苹果在研…
研究前沿 爱范儿 · 11-9 阅读 18·访客 18
一篇读懂多模态统一训练:视觉、语言、音频的联合建模前沿
从 CLIP 对比预训练到连接器路线与原生 token 统一路线:图文音频联合建模的架构分歧、训练稳定性问题、模态不平衡与数据配比,一篇讲透。
大模型 精选 · 原创 · 11-9 阅读 68·访客 67
让模型看懂图表与几何:视觉推理研究的演进与瓶颈
从 ChartQA 的图表问答到 MathVista 的图文数学推理,再到「视觉盲」测试暴露的底层缺陷,梳理视觉推理评测五年演进与当前真正的技术瓶颈。
研究前沿 精选 · 原创 · 11-9 阅读 56·访客 56
具身智能研究:从仿真到实体的机器人学习路径
从域随机化、GPU 并行仿真到 VLA 大模型与真机世界模型,梳理具身智能从仿真走向实体机器人的完整学习路径、代表论文与开放问题。
研究前沿 精选 · 原创 · 11-9 阅读 61·访客 61
MotionGraphicsBench: Evaluating Motion Graphics Generation with Programmatic Verification
Large language models can generate motion graphics from static images and natural language requests, but it remains diff…
行业动态 HuggingFace Daily Papers · 11-7 阅读 4·访客 4
Subliminal Learning in Vision-Language Models
Subliminal learning is a phenomenon in large language models (LLMs) where a teacher passes a hidden preference to a stud…
大模型 HuggingFace Daily Papers · 11-4 阅读 4·访客 4
多模态越狱:藏在图片与音频里的新攻击面
给模型装上眼睛和耳朵,等于给它开了一扇绕过安全训练的新门:把恶意指令藏进图片、用高频音频注入、在图表里编码攻击文本——多模态模型的攻击面远大于纯文本。本文讲清多模态越狱的主要技术路线、为什么它们比文本越狱更难防、以及防御的纵深设计。
研究前沿 精选 · 原创 · 11-3 阅读 73·访客 73
NVIDIA Releases Nemotron 3.5 Bolt E4B 4GB: An Open Small Language Model for Local Voice Agents, Game Characters and Robots
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体 MarkTechPost · 10-24 阅读 84·访客 82
WeVisOne-VL: Unified Visual Perception from Objects to Pixels
Existing vision-language models (VLMs) remain unreliable for fine-grained visual perception. In this report, we present …
行业动态 HuggingFace Daily Papers · 10-22 阅读 18·访客 18
Bringing the Spark to Europe: Local Intelligence Powers Next Wave of Agents
At GTC Berlin, NVIDIA is showcasing its full lineup of local AI hardware. RTX Spark, which launched last week, powers pe…
智能体 NVIDIA Blog · 10-21 阅读 79·访客 79
MambaAct: Efficient Mamba Action Experts for Flow-Matching Vision-Language-Action Models
Vision-language-action (VLA) models have shown strong promise for generalist robotic manipulation, but their inference c…
智能体 HuggingFace Daily Papers · 10-16 阅读 51·访客 51
清华具身模型登顶全球第一!突围GPT-6、英伟达,不靠外挂和额外数据
田, 晏林* 2026-10-09 11:52:21 来源:量子位 星动纪元选择将视频预测与动作学习分阶段训练,重点不是「视频、动作一锅炖」,而是把两者「解耦」,重新「排序」。 田晏林 发自 凹非寺 量子位 | 公众号 QbitAI 好家…
大模型 量子位 · 10-9 阅读 53·访客 53
Embodied Turing Machines: Stateful Code for Robot Recursive Self-Improvement
Most robot policies keep a model in the control loop: a VLA maps observations to actions, and an Agent Harness, such as …
智能体 HuggingFace Daily Papers · 10-8 阅读 82·访客 81
AI breakthroughs in robotics won’t change your life any time soon
The story is a collaboration between MIT Technology Review and Aventine, a non-profit research foundation that creates a…
智能体 MIT Technology Review · 10-8 阅读 97·访客 97
BudgetPix: Compute-Adaptive Tokenization for Pixel-Space Image Diffusion
Most image generation models rely on uniform tokenization, allocating the exact same computational budget to equally-siz…
行业动态 HuggingFace Daily Papers · 10-8 阅读 27·访客 27
VLA(视觉-语言-动作)模型进化史:机器人如何「看懂再动手」
从 RT-2 把机器人动作当文本 token 输出,到 OpenVLA 开源 7B、π0 用流匹配跑到 50Hz、GR00T 与 Helix 转向双系统架构——本文梳理 VLA 模型三年的演进脉络:动作怎么表示、参数怎么变小、控制频率怎么上去,以及开源与闭源两条路线的分野。
研究前沿 精选 · 原创 · 10-7 阅读 700·访客 678
多模态 RAG:当文档里的关键信息藏在表格与图表里
真实企业文档的多数信息不在正文段落里,而在表格、图表与版式关系里——文本 RAG 的 OCR 管线在这里系统性失真。本文拆解两条补齐路线(解析增强 vs ColPali 式视觉检索)与生成端的图文编排,给出选型权衡。
后端技术 精选 · 原创 · 10-7 阅读 540·访客 537
When to Rethink: Learning Multi-Perspective Self-Verification for Vision-Language Models
Vision-language models (VLMs) have achieved strong performance in multimodal reasoning, yet they remain prone to generat…
研究前沿 HuggingFace Daily Papers · 10-4 阅读 29·访客 29
VIEScore2:带空间定位解释的统一图像评估模型
VIEScore2 将图像表示为 N×N 网格,在单次前向中同时预测图像质量分数与缺陷位置,并在 38K 样本上结合监督微调与 GRPO 训练。
研究前沿 HuggingFace Daily Papers · 10-1 阅读 35·访客 35
刚刚,GPT-6 Astra接上宇树G1,把厨房收拾了!
henry* 2026-09-30 15:54:54 来源:量子位 让GPT把机器人技能当工具调用 henry 发自 凹非寺 量子位 | 公众号 QbitAI 机器人的GPT时刻,还真得靠GPT(doge)! 刚刚,GPT-6 Astra…
智能体 量子位 · 9-30 阅读 121·访客 119
工业创新进入“组队局”,拆解西门子Xcelerator开放生态的赋能链路
田, 晏林* 2026-09-28 20:43:53 来源:量子位 工业平台已经卷到帮伙伴拿线索、做Agent、出海了 田晏林 发自 凹非寺 量子位 | 公众号 QbitAI 一家技术公司想把一个工业创新方案卖进工厂,中间大概要过多少关?…
智能体 量子位 · 9-28 阅读 118·访客 117
Google Research Introduces an AI Video Co-Director: 4 Agentic Frameworks for Coherent, Minutes-Long Video Generation
Google Research has introduced an **AI video co-director** for long-form video generation. The suite of 4 agentic framew…
智能体 MarkTechPost · 9-28 阅读 128·访客 127
Researchers plug GPT-6 Astra directly into a robot and let it clean up an unfamiliar kitchen
Sep 27, 2026 Researchers from Stanford and Caltech have built HomeBody,** a system that lets a Unitree G1 robot autonomo…
智能体 The Decoder · 9-27 阅读 117·访客 114
Liquid AI Releases LFM2.5-VL-3B-DSpark: Speculative Decoding for Vision-Language Models With Up to 3.13x Faster Decoding
Liquid AI has announced LFM2.5-VL-3B-DSpark, an experimental speculative-decoding draft model for its LFM2.5-VL-3B visio…
行业动态 MarkTechPost · 9-26 阅读 48·访客 48
AI开始研究Physical AI:FSD级团队亮出首版模型Simate-beta,空降RoboDojo
田, 晏林* 2026-09-26 17:07:41 来源:量子位 Simate将训练、推理与评测全流程接入自研Infra,通过极致的任务编排与资源调度,同时并行推进数十条相互独立的研究路线。 Jay 发自 凹非寺 量子位 | 公众号 Q…
研究前沿 量子位 · 9-26 阅读 52·访客 52
InternW0-Δ: A World Action Model Bridging Predictive Dynamics and Actions with 20K+ Hours of Open Data
World Action Models (WAMs) jointly model visual dynamics and action generation for generalist robot manipulation. A cent…
智能体 HuggingFace Daily Papers · 9-25 阅读 73·访客 73
早报|GPT-6 Sol发布,价格腰斩/特努斯:Siri AI不应代替人际关系/4999起,OPPO Find X10系列发布
🧑‍💻GPT-6 Sol 与 Luna 发布,API 价格降低 50% 🧠Claude Opus 5.5 发布,典型任务运行成本降低 40% 🍎特努斯:Siri AI 不应代替人际关系 📱OPPO Find X10 系列发布:49…
大模型 爱范儿 · 9-23 阅读 63·访客 63
HARMONY: Hierarchical Agentic Reasoning for MONocular Image-to-Scene Synthesis
Compositional 3D scene reconstruction has recently been explored from two directions: agentic reasoning that provides se…
智能体 HuggingFace Daily Papers · 9-22 阅读 93·访客 92
RoboFollow: Unveiling the Instruction Following Mirage in Embodied Agents
Modern embodied agents achieve impressive success rates, yet their actual instruction-following ability is far weaker th…
智能体 HuggingFace Daily Papers · 9-22 阅读 80·访客 80
X-Planner: Event-Structured Task Planning for Embodied Intelligence
Task planning bridges high-level instructions and executable behavior in long-horizon manipulation, yet modern Vision-La…
行业动态 HuggingFace Daily Papers · 9-21 阅读 27·访客 27
RULER: Instance-aware Rubric Rewards for SVG Generation
Generating Scalable Vector Graphics (SVG) code from natural-language instructions is an open-ended task without absolute…
智能体 HuggingFace Daily Papers · 9-21 阅读 76·访客 75
All-in-One Multilingual Scene Text Recognition with Script-aware Mixture-of-Experts
Multilingual scene text recognition (STR) remains challenging due to the scarcity of training data for most languages an…
行业动态 HuggingFace Daily Papers · 9-21 阅读 35·访客 35
PackLab: A Comprehensive Framework for Developing, Training, and Evaluating MLLMs in Robotic Bin Packing
Robotic bin packing requires long-horizon sequential decision-making, as each object placement affects the available spa…
智能体 HuggingFace Daily Papers · 9-20 阅读 89·访客 88
Transferring the Intelligence of VLMs to Robotic Control
Humans can seamlessly adapt to both physical and digital worlds, suggesting that while a digital-to-real gap exists in e…
智能体 HuggingFace Daily Papers · 9-19 阅读 73·访客 72
Paint-Anything: Unified Any-Color Control for Image Generation and Editing
Professional design requires any-color control: the ability to specify an object's target color with any 24-bit hex valu…
行业动态 HuggingFace Daily Papers · 9-17 阅读 44·访客 44