{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-07-23T06:40:50.084Z","headline":"ABot-World-0：单张桌面级GPU实现无限交互式世界生成","description":"ABot-World-0是一个动作条件视频世界模型，能在单张NVIDIA RTX 5090 GPU上以720P分辨率、最高16 FPS、1.2秒动作到首帧延迟和约19 GiB峰值显存预算运行无限交互式世界生成。该模型采用统一的帧同步键盘动作接口，通过LongForcing分布匹配阶段解决长程自回归漂移问题，并集成轻量级VAE解码器、低比特DiT推理和局部上下文KV缓存等系统级优化。","url":"https://www.aioga.com/news/cmrvls9qg03jtbihbfhtx203y/","mainEntityOfPage":"https://www.aioga.com/news/cmrvls9qg03jtbihbfhtx203y/","datePublished":"2026-07-21T15:26:50.000Z","dateModified":"2026-07-21T15:26:50.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://arxiv.org/abs/2607.19191","https://aihot.virxact.com/items/cmrvls9qg03jtbihbfhtx203y"],"canonicalUrl":"https://www.aioga.com/news/cmrvls9qg03jtbihbfhtx203y/","directAnswer":{"@type":"Answer","text":"ABot-World-0是一种动作条件视频世界模型。据摘要，其可在单张NVIDIA RTX 5090上以720P、最高16 FPS运行，并支持无限交互式世界生成。","url":"https://www.aioga.com/news/cmrvls9qg03jtbihbfhtx203y/","dateCreated":"2026-07-21T15:26:50.000Z","author":{"@type":"Organization","@id":"https://www.aioga.com/authors/aioga-editorial/#editorial-team","name":"Aioga Editorial Team","url":"https://www.aioga.com/authors/aioga-editorial/"}},"evidence":[{"@type":"CreativeWork","name":"arXiv source article","url":"https://arxiv.org/abs/2607.19191","datePublished":"2026-07-21T15:26:50.000Z","provider":{"@type":"Organization","name":"arXiv","url":"https://arxiv.org/abs/2607.19191"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.virxact.com/items/cmrvls9qg03jtbihbfhtx203y","datePublished":"2026-07-21T15:26:50.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.virxact.com/items/cmrvls9qg03jtbihbfhtx203y"}}],"aggregationSource":"HuggingFace Daily Papers（社区热门论文）","originalPublisher":{"name":"arXiv","url":"https://arxiv.org/abs/2607.19191"},"article":{"id":"cmrvls9qg03jtbihbfhtx203y","slug":"cmrvls9qg03jtbihbfhtx203y","url":"https://www.aioga.com/news/cmrvls9qg03jtbihbfhtx203y/","title":"ABot-World-0：单张桌面级GPU实现无限交互式世界生成","title_en":"ABot-World-0： Infinite Interactive World Rollout on a Single Desktop GPU","summary":"ABot-World-0是一个动作条件视频世界模型，能在单张NVIDIA RTX 5090 GPU上以720P分辨率、最高16 FPS、1.2秒动作到首帧延迟和约19 GiB峰值显存预算运行无限交互式世界生成。该模型采用统一的帧同步键盘动作接口，通过LongForcing分布匹配阶段解决长程自回归漂移问题，并集成轻量级VAE解码器、低比特DiT推理和局部上下文KV缓存等系统级优化。","source":"HuggingFace Daily Papers（社区热门论文）","sourceUrl":"https://arxiv.org/abs/2607.19191","aiHotUrl":"https://aihot.virxact.com/items/cmrvls9qg03jtbihbfhtx203y","publishedAt":"2026-07-21T15:26:50.000Z","category":"论文研究","score":83,"selected":true,"articleBody":["arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.","Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs ：https://info.arxiv.org/labs/index.html."],"articleImages":[{"sourceUrl":"https://arxiv.org/static/base/1.0.1/images/funders/simons-foundation.png","alt":"Simons Foundation","afterParagraph":1,"url":"/media/articles/cmrvls9qg03jtbihbfhtx203y/e2d7f38d62f5ca91.png"},{"sourceUrl":"https://arxiv.org/static/base/1.0.1/images/funders/simons-foundation-international.png","alt":"Simons Foundation International","afterParagraph":1,"url":"/media/articles/cmrvls9qg03jtbihbfhtx203y/1d56e29c5557cbdc.png"},{"sourceUrl":"https://arxiv.org/static/base/1.0.1/images/funders/schmidt-sciences.png","alt":"Schmidt Sciences","afterParagraph":1,"url":"/media/articles/cmrvls9qg03jtbihbfhtx203y/8e18212b8219c104.png"}],"mediaStatus":"ok","articleBodyZh":["arXivLabs 是一个框架，允许合作者直接在我们的网站上开发和分享新的 arXiv 功能。","有一个可以为 arXiv 社区增加价值的项目想法吗？了解更多关于 arXivLabs 的信息：https://info.arxiv.org/labs/index.html。"],"translationStatus":"translated","bodyOrigin":"source-page","editorial":{"summary":"ABot-World-0是一种动作条件视频世界模型。据摘要，其可在单张NVIDIA RTX 5090上以720P、最高16 FPS运行，并支持无限交互式世界生成。","background":"该模型采用统一的帧同步键盘动作接口，并通过LongForcing分布匹配阶段处理长程自回归漂移，同时结合轻量级VAE解码器等系统优化。","viewpoint":"Aioga判断，这项研究值得关注之处在于将交互生成与桌面级单卡推理结合，但现有材料未提供画质、稳定性、任务覆盖范围或对比实验细节。","implications":"若摘要所述性能在更多场景中得到验证，该方案可能降低交互式世界模型的运行门槛，并为低比特DiT推理和局部上下文KV缓存的组合应用提供参考。","nextStep":"值得关注论文后续披露的完整实验设置、长时间交互结果、画质评估及硬件测试方法，并核验最高16 FPS、1.2秒延迟和约19 GiB显存数据的适用条件。","evidenceRefs":["title","summary","articleBody","source"],"status":"published","aiGenerated":true,"autoApproved":true,"generatedBy":"aioga-editorial:gpt-5.6-sol","reviewedBy":"aioga-editorial-review:gpt-5.6-sol","generatedAt":"2026-07-22T18:14:26.651Z","sourceHash":"9941c2639bdac7e4","review":{"approved":true,"groundedness":97,"clarity":94,"duplicationRisk":18,"blockingIssues":[],"notes":["候选内容中的模型类型、硬件、分辨率、帧率、动作接口、LongForcing及系统优化均与来源摘要一致。","对运行门槛和技术参考价值的判断使用了“若”“可能”等条件性措辞，没有将尚未验证的影响表述为既成事实。","viewpoint明确标注为“Aioga判断”，并如实指出现有材料缺少画质、稳定性、任务覆盖范围和对比实验细节。","nextStep要求核验性能数据的适用条件，合理反映了来源仅提供摘要信息的证据局限。"]},"validation":{"passed":true,"mode":"ai-auto","revisions":0,"checks":["schema","length","source-attribution","low-source-overlap","no-html","independent-ai-review"]}},"tags":["论文研究","HuggingFace Daily Papers（社区热门论文）"],"translations":{"zh-CN":{"title":"ABot-World-0：单张桌面级GPU实现无限交互式世界生成","summary":"ABot-World-0是一个动作条件视频世界模型，能在单张NVIDIA RTX 5090 GPU上以720P分辨率、最高16 FPS、1.2秒动作到首帧延迟和约19 GiB峰值显存预算运行无限交互式世界生成。该模型采用统一的帧同步键盘动作接口，通过LongForcing分布匹配阶段解决长程自回归漂移问题，并集成轻量级VAE解码器、低比特DiT推理和局部上下文KV缓存等系统级优化。","category":"论文研究","source":"arXiv","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0：单张桌面级GPU实现无限交互式世界生成 - Aioga AI资讯","description":"ABot-World-0是一个动作条件视频世界模型，能在单张NVIDIA RTX 5090 GPU上以720P分辨率、最高16 FPS、1.2秒动作到首帧延迟和约19 GiB峰值显存预算运行无限交互式世界生成。该模型采用统一的帧同步键盘动作接口，通过LongForcing分布匹配阶段解决长程自回归漂移问题，并集成轻量级VAE解码器、低比特DiT推理和局部上下...","url":"https://www.aioga.com/news/cmrvls9qg03jtbihbfhtx203y/"},"en":{"title":"ABot-World-0: Creating an infinite interactive world with a single desktop-grade GPU","summary":"ABot-World-0 is an action-conditioned video world model that can run continuous interactive world generation on a single NVIDIA RTX 5090 GPU at 720P resolution, up to 16 FPS, with 1.2-second action-to-first-frame latency, and approximately 19 GiB peak memory usage. The model uses a unified frame-synchronized keyboard action interface, addresses long-term autoregressive drift issues through a LongForcing distribution matching phase, and integrates system-level optimizations such as a lightweight VAE decoder, low-bit DiT inference, and local context KV caching.","category":"Research","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: Creating an infinite interactive world with a single desktop-grade GPU - Aioga AI News","description":"ABot-World-0 is an action-conditioned video world model that can run continuous interactive world generation on a single NVIDIA RTX 5090 GPU at 720P resolution, up to 16 FPS, with...","url":"https://www.aioga.com/en/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:45:57.879Z"},"ja":{"title":"ABot-World-0：単一のデスクトップGPUで無限のインタラクティブワールド生成を実現","summary":"ABot-World-0は、アクション条件付きビデオワールドモデルで、単一のNVIDIA RTX 5090 GPU上で720P解像度、最高16 FPS、1.2秒のアクションから最初のフレームへの遅延、約19 GiBのピークメモリ消費で無限にインタラクティブなワールド生成を実行できます。このモデルは統一されたフレーム同期キーボードアクションインターフェースを採用し、LongForcing分布マッチング段階を通じて長期自己回帰のドリフト問題を解決し、軽量VAEデコーダー、低ビットDiT推論、局所コンテキストKVキャッシュなどのシステムレベルの最適化を統合しています。","category":"論文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0：単一のデスクトップGPUで無限のインタラクティブワールド生成を実現 - Aioga AIニュース","description":"ABot-World-0は、アクション条件付きビデオワールドモデルで、単一のNVIDIA RTX 5090 GPU上で720P解像度、最高16 FPS、1.2秒のアクションから最初のフレームへの遅延、約19 GiBのピークメモリ消費で無限にインタラクティブなワールド生成を実行できます。このモデルは統一されたフレーム同期キーボードアクションインターフェースを採...","url":"https://www.aioga.com/ja/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:46:09.165Z"},"ko":{"title":"ABot-World-0: 단일 데스크톱급 GPU로 무한 상호작용 세계 생성 구현","summary":"ABot-World-0는 단일 NVIDIA RTX 5090 GPU에서 720P 해상도, 최대 16 FPS, 1.2초 동작-첫 프레임 지연과 약 19 GiB 피크 VRAM 예산으로 무한 상호작용 세계 생성이 가능한 액션 조건 비디오 월드 모델입니다. 이 모델은 통합 프레임 동기 키보드 액션 인터페이스를 사용하며, LongForcing 분포 매칭 단계를 통해 장기 자기회귀 편차 문제를 해결하고, 경량 VAE 디코더, 저비트 DiT 추론 및 로컬 컨텍스트 KV 캐시 등 시스템 수준 최적화를 통합합니다.","category":"연구","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: 단일 데스크톱급 GPU로 무한 상호작용 세계 생성 구현 - Aioga AI 뉴스","description":"ABot-World-0는 단일 NVIDIA RTX 5090 GPU에서 720P 해상도, 최대 16 FPS, 1.2초 동작-첫 프레임 지연과 약 19 GiB 피크 VRAM 예산으로 무한 상호작용 세계 생성이 가능한 액션 조건 비디오 월드 모델입니다. 이 모델은 통합 프레임 동기 키보드 액션 인터페이스를 사용하며, LongF...","url":"https://www.aioga.com/ko/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:46:57.298Z"},"es":{"title":"ABot-World-0: Generación de mundos interactivos infinitos con una sola GPU de escritorio","summary":"ABot-World-0 es un modelo de mundo de video con acciones condicionales, capaz de ejecutar generación de mundos interactivos ilimitados en una sola GPU NVIDIA RTX 5090 a una resolución de 720P, con un máximo de 16 FPS, un retraso de acción a primer cuadro de 1.2 segundos y un presupuesto de memoria gráfica pico de aproximadamente 19 GiB. El modelo utiliza una interfaz de acción de teclado sincronizada por fotogramas unificada, resuelve el problema de deriva autoregresiva a largo plazo mediante la fase de coincidencia de distribución LongForcing, y está integrado con optimizaciones a nivel de sistema como un decodificador VAE ligero, inferencia DiT de baja cantidad de bits y almacenamiento en caché KV de contexto local.","category":"Investigación","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: Generación de mundos interactivos infinitos con una sola GPU de escritorio - Aioga Noticias de IA","description":"ABot-World-0 es un modelo de mundo de video con acciones condicionales, capaz de ejecutar generación de mundos interactivos ilimitados en una sola GPU NVIDIA RTX 5090 a una resoluc...","url":"https://www.aioga.com/es/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:46:55.838Z"},"fr":{"title":"ABot-World-0 : Génération de monde interactif infini avec une seule GPU de bureau","summary":"ABot-World-0 est un modèle de monde vidéo conditionné par l'action, capable de générer un monde interactif illimité sur un seul GPU NVIDIA RTX 5090 avec une résolution de 720P, jusqu'à 16 FPS, un délai de 1,2 seconde entre l'action et la première image, et un budget de mémoire maximale d'environ 19 GiB. Le modèle utilise une interface d'action clavier synchronisée avec les images, résout le problème de dérive auto-régressive à long terme via une phase de correspondance distribuée LongForcing, et intègre des optimisations au niveau système telles qu'un décodeur VAE léger, une inférence DiT à faible débit binaire et un cache KV local contextuel.","category":"Recherche","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0 : Génération de monde interactif infini avec une seule GPU de bureau - Aioga Actualités IA","description":"ABot-World-0 est un modèle de monde vidéo conditionné par l'action, capable de générer un monde interactif illimité sur un seul GPU NVIDIA RTX 5090 avec une résolution de 720P, jus...","url":"https://www.aioga.com/fr/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:47:47.067Z"},"de":{"title":"ABot-World-0: Einzelne Desktop-GPU ermöglicht unbegrenzte interaktive Welterzeugung","summary":"ABot-World-0 ist ein Aktions-basiertes Videoweltmodell, das auf einer einzelnen NVIDIA RTX 5090 GPU bei 720P-Auflösung, maximal 16 FPS, einer Aktions-zu-Erstbild-Verzögerung von 1,2 Sekunden und einem Spitzen-Videospeicherverbrauch von etwa 19 GiB eine unbegrenzte interaktive Weltgenerierung ausführen kann. Das Modell verwendet eine einheitliche, rahmensynchrone Tastaturaktionsschnittstelle, löst das Problem der langfristigen autoregressiven Drift durch eine LongForcing-Verteilungsabstimmungsphase und integriert Systemoptimierungen wie einen leichten VAE-Dekodierer, Low-Bit-DiT-Inferenz und lokale Kontext-KV-Caches.","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: Einzelne Desktop-GPU ermöglicht unbegrenzte interaktive Welterzeugung - Aioga KI-News","description":"ABot-World-0 ist ein Aktions-basiertes Videoweltmodell, das auf einer einzelnen NVIDIA RTX 5090 GPU bei 720P-Auflösung, maximal 16 FPS, einer Aktions-zu-Erstbild-Verzögerung von 1,...","url":"https://www.aioga.com/de/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:47:47.182Z"},"pt-BR":{"title":"ABot-World-0: Geração de mundo interativo infinito com uma única GPU de nível desktop","summary":"ABot-World-0 é um modelo de mundo em vídeo condicional por ação, capaz de gerar mundos interativos ilimitados com resolução 720P, até 16 FPS, 1,2 segundos de atraso da ação para o primeiro quadro e cerca de 19 GiB de pico de memória de vídeo em uma única GPU NVIDIA RTX 5090. O modelo utiliza uma interface unificada de ações de teclado sincronizadas por quadro, resolve o problema de deriva autoregressiva de longo alcance através da fase de correspondência de distribuição LongForcing, e integra otimizações de sistema como decodificador VAE leve, inferência DiT de baixa taxa de bits e cache KV de contexto local.","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: Geração de mundo interativo infinito com uma única GPU de nível desktop - Aioga Notícias de IA","description":"ABot-World-0 é um modelo de mundo em vídeo condicional por ação, capaz de gerar mundos interativos ilimitados com resolução 720P, até 16 FPS, 1,2 segundos de atraso da ação para o...","url":"https://www.aioga.com/pt-BR/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:48:33.152Z"},"ru":{"title":"ABot-World-0: Реализация бесконечного интерактивного генератора миров на одном настольном GPU","summary":"ABot-World-0 — это модель видео-мира с условием действий, способная работать на одной GPU NVIDIA RTX 5090 с разрешением 720P, с максимальной частотой 16 кадров в секунду, задержкой от действия до первого кадра 1,2 секунды и пиковым потреблением видеопамяти около 19 ГиБ для генерации интерактивного мира без ограничений. Модель использует унифицированный интерфейс синхронизации кадров для клавиатурных действий, решает проблему дрейфа длинной автокорреляции на этапе сопоставления распределений LongForcing и интегрирует оптимизации на системном уровне, такие как легковесный VAE-декодер, низкобитное вывод DiT и локальный кэш KV контекста.","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: Реализация бесконечного интерактивного генератора миров на одном настольном GPU - Aioga Новости ИИ","description":"ABot-World-0 — это модель видео-мира с условием действий, способная работать на одной GPU NVIDIA RTX 5090 с разрешением 720P, с максимальной частотой 16 кадров в секунду, задержкой...","url":"https://www.aioga.com/ru/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:48:33.051Z"},"ar":{"title":"ABot-World-0: إنشاء عالم تفاعلي غير محدود باستخدام بطاقة رسومات واحدة من فئة سطح المكتب","summary":"ABot-World-0 هو نموذج عالم فيديو يعتمد على إجراء معين، يمكنه تشغيل توليد عالم تفاعلي لا نهائي بدقة 720P وسرعة تصل إلى 16 إطارًا في الثانية، مع تأخير 1.2 ثانية من الإجراءات إلى الإطار الأول وميزانية ذروة ذاكرة تقريبية تبلغ 19 GiB على بطاقة NVIDIA RTX 5090 GPU واحدة. يستخدم النموذج واجهة إجراءات لوحة مفاتيح متزامنة مع الإطارات موحدة، ويحل مشكلة الانحراف الذاتي طويل المدى من خلال مرحلة مطابقة التوزيع LongForcing، ويدمج تحسينات على مستوى النظام مثل محول VAE خفيف، واستنتاج DiT منخفض البت، وذاكرة مؤقتة KV للسياق المحلي.","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: إنشاء عالم تفاعلي غير محدود باستخدام بطاقة رسومات واحدة من فئة سطح المكتب - Aioga أخبار الذكاء الاصطناعي","description":"ABot-World-0 هو نموذج عالم فيديو يعتمد على إجراء معين، يمكنه تشغيل توليد عالم تفاعلي لا نهائي بدقة 720P وسرعة تصل إلى 16 إطارًا في الثانية، مع تأخير 1.2 ثانية من الإجراءات إلى الإط...","url":"https://www.aioga.com/ar/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:49:22.430Z"},"hi":{"title":"ABot-World-0: एक सिंगल डेस्कटॉप GPU के साथ अनंत इंटरैक्टिव वर्ल्ड जनरेशन संभव","summary":"ABot-World-0 एक एक्शन-कंडीशनल वीडियो वर्ल्ड मॉडल है, जो एकल NVIDIA RTX 5090 GPU पर 720P रिज़ॉल्यूशन, अधिकतम 16 FPS, 1.2 सेकंड एक्शन से पहले फ़्रेम डिले और लगभग 19 GiB पीक मेमोरी बजट के साथ अनिश्चित इंटरैक्टिव वर्ल्ड जेनरेशन चला सकता है। यह मॉडल एक एकीकृत फ़्रेम-सिंक कीबोर्ड एक्शन इंटरफ़ेस का उपयोग करता है, LongForcing वितरण मिलान चरण के माध्यम से लंबी अवधि की ऑटोरेग्रेसीव ड्रिफ्ट समस्या को हल करता है, और इसमें हल्के VAE डिकोडर, लो-बिट DiT इन्फ़रेंस और स्थानीय कॉन्टेक्स्ट KV कैश जैसी सिस्टम-स्तरीय अनुकूलन शामिल हैं।","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: एक सिंगल डेस्कटॉप GPU के साथ अनंत इंटरैक्टिव वर्ल्ड जनरेशन संभव - Aioga AI समाचार","description":"ABot-World-0 एक एक्शन-कंडीशनल वीडियो वर्ल्ड मॉडल है, जो एकल NVIDIA RTX 5090 GPU पर 720P रिज़ॉल्यूशन, अधिकतम 16 FPS, 1.2 सेकंड एक्शन से पहले फ़्रेम डिले और लगभग 19 GiB पीक मेमोरी बज...","url":"https://www.aioga.com/hi/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:49:21.473Z"},"it":{"title":"ABot-World-0: Generazione di mondi interattivi illimitata con una singola GPU desktop","summary":"ABot-World-0 è un modello video del mondo con azioni condizionate, in grado di far girare la generazione di mondi interattivi illimitati su una singola GPU NVIDIA RTX 5090 a risoluzione 720P, con un massimo di 16 FPS, un ritardo di azione al primo frame di 1,2 secondi e un picco di memoria di circa 19 GiB. Il modello utilizza un'interfaccia di azione da tastiera sincronizzata con i frame unificata, risolve il problema della deriva autoregressiva a lungo raggio attraverso una fase di adattamento distribuito LongForcing e integra ottimizzazioni a livello di sistema come un decoder VAE leggero, inferenza DiT a basso bit e cache KV del contesto locale.","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: Generazione di mondi interattivi illimitata con una singola GPU desktop - Aioga Notizie IA","description":"ABot-World-0 è un modello video del mondo con azioni condizionate, in grado di far girare la generazione di mondi interattivi illimitati su una singola GPU NVIDIA RTX 5090 a risolu...","url":"https://www.aioga.com/it/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:50:13.541Z"},"nl":{"title":"ABot-World-0: Een enkele desktop-GPU realiseert onbeperkte interactieve wereldgeneratie","summary":"ABot-World-0 is een actievoorwaardelijk videowereldmodel dat op een enkele NVIDIA RTX 5090 GPU kan draaien met een resolutie van 720P, tot 16 FPS, 1,2 seconden actietot-eerste-frame vertraging en een piekgeheugenbudget van ongeveer 19 GiB voor oneindige interactieve wereldgeneratie. Het model maakt gebruik van een uniform frame-synchronisatie toetsenbordactie-interface, lost het lange-afstands autoregressieve driftprobleem op via een LongForcing distributiematchfase, en integreert systeemniveauoptimalisaties zoals een lichtgewicht VAE-decoder, laagbit DiT-inferentie en lokale context KV-cache.","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: Een enkele desktop-GPU realiseert onbeperkte interactieve wereldgeneratie - Aioga AI-nieuws","description":"ABot-World-0 is een actievoorwaardelijk videowereldmodel dat op een enkele NVIDIA RTX 5090 GPU kan draaien met een resolutie van 720P, tot 16 FPS, 1,2 seconden actietot-eerste-fram...","url":"https://www.aioga.com/nl/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:50:08.059Z"},"tr":{"title":"ABot-World-0: Tek bir masaüstü GPU ile sonsuz etkileşimli dünya üretimi","summary":"ABot-World-0, tek bir NVIDIA RTX 5090 GPU üzerinde 720P çözünürlükte, maksimum 16 FPS hızında, 1,2 saniye hareketten ilk kare gecikmesiyle ve yaklaşık 19 GiB tepe bellek bütçesi ile sınırsız etkileşimli dünya üretimi sağlayabilen bir aksiyon tabanlı koşullu video dünya modelidir. Model, birleştirilmiş kare senkronize klavye hareket arayüzü kullanır, uzun süreli oto-regresif kayma sorununu LongForcing dağılım eşleme aşaması ile çözer ve hafif VAE çözücü, düşük bit DiT çıkarımı ve yerel bağlam KV önbelleği gibi sistem düzeyinde optimizasyonları entegre eder.","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: Tek bir masaüstü GPU ile sonsuz etkileşimli dünya üretimi - Aioga AI Haberleri","description":"ABot-World-0, tek bir NVIDIA RTX 5090 GPU üzerinde 720P çözünürlükte, maksimum 16 FPS hızında, 1,2 saniye hareketten ilk kare gecikmesiyle ve yaklaşık 19 GiB tepe bellek bütçesi il...","url":"https://www.aioga.com/tr/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:51:09.619Z"},"vi":{"title":"ABot-World-0: Tạo thế giới tương tác vô hạn bằng một GPU cấp máy tính để bàn","summary":"ABot-World-0 là một mô hình thế giới video điều kiện hành động, có thể chạy tạo thế giới tương tác vô hạn với độ phân giải 720P, tốc độ tối đa 16 FPS, độ trễ từ hành động đến khung đầu 1,2 giây và khoảng 19 GiB bộ nhớ đỉnh trên một GPU NVIDIA RTX 5090 đơn. Mô hình này sử dụng giao diện hành động bàn phím đồng bộ khung thống nhất, giải quyết vấn đề trôi tự hồi quy dài hạn thông qua giai đoạn khớp phân phối LongForcing, và tích hợp các tối ưu hóa cấp hệ thống như bộ giải mã VAE nhẹ, suy luận DiT bit thấp và bộ đệm KV ngữ cảnh cục bộ.","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: Tạo thế giới tương tác vô hạn bằng một GPU cấp máy tính để bàn - Tin tức AI Aioga","description":"ABot-World-0 là một mô hình thế giới video điều kiện hành động, có thể chạy tạo thế giới tương tác vô hạn với độ phân giải 720P, tốc độ tối đa 16 FPS, độ trễ từ hành động đến khung...","url":"https://www.aioga.com/vi/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:51:05.878Z"},"id":{"title":"ABot-World-0: GPU desktop tunggal mewujudkan generasi dunia interaktif tak terbatas","summary":"ABot-World-0 adalah model dunia video berbasis aksi bersyarat, yang dapat menjalankan generasi dunia interaktif tak terbatas pada resolusi 720P dengan maksimum 16 FPS, 1,2 detik latensi dari aksi ke frame pertama, dan anggaran memori puncak sekitar 19 GiB pada satu GPU NVIDIA RTX 5090. Model ini menggunakan antarmuka aksi keyboard sinkron frame yang seragam, menyelesaikan masalah drift autoregresif jangka panjang melalui tahap pencocokan distribusi LongForcing, dan mengintegrasikan optimisasi sistem seperti decoder VAE ringan, inference DiT berbit rendah, dan cache KV konteks lokal.","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: GPU desktop tunggal mewujudkan generasi dunia interaktif tak terbatas - Berita AI Aioga","description":"ABot-World-0 adalah model dunia video berbasis aksi bersyarat, yang dapat menjalankan generasi dunia interaktif tak terbatas pada resolusi 720P dengan maksimum 16 FPS, 1,2 detik la...","url":"https://www.aioga.com/id/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:51:54.501Z"},"th":{"title":"ABot-World-0: การสร้างโลกเชิงโต้ตอบแบบไม่จำกัดด้วย GPU ระดับเดสก์ท็อปเพียงตัวเดียว","summary":"ABot-World-0 คือโมเดลโลกวิดีโอแบบมีเงื่อนไขการเคลื่อนไหว สามารถรันการสร้างโลกแบบอินเทอร์แอคทีฟไม่จำกัดบนการ์ดจอ NVIDIA RTX 5090 เพียงใบเดียว ด้วยความละเอียด 720P สูงสุด 16 FPS ความล่าช้าจากการกระทำถึงเฟรมแรก 1.2 วินาที และใช้หน่วยความจำสูงสุดประมาณ 19 GiB โมเดลนี้ใช้อินเทอร์เฟซการกระทำผ่านแป้นพิมพ์แบบซิงโครไนซ์เฟรมเดียว โดยแก้ไขปัญหาการเบี่ยงเบนอัตโนมัติระยะยาวด้วยขั้นตอนการจับคู่การกระจาย LongForcing และรวมระบบปรับแต่งระดับซิสเต็ม เช่น ตัวถอดรหัส VAE น้ำหนักเบา การอนุมาน DiT แบบบิตต่ำ และแคช KV บริบทท้องถิ่น","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: การสร้างโลกเชิงโต้ตอบแบบไม่จำกัดด้วย GPU ระดับเดสก์ท็อปเพียงตัวเดียว - ข่าว AI Aioga","description":"ABot-World-0 คือโมเดลโลกวิดีโอแบบมีเงื่อนไขการเคลื่อนไหว สามารถรันการสร้างโลกแบบอินเทอร์แอคทีฟไม่จำกัดบนการ์ดจอ NVIDIA RTX 5090 เพียงใบเดียว ด้วยความละเอียด 720P สูงสุด 16 FPS ความ...","url":"https://www.aioga.com/th/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:52:03.582Z"},"pl":{"title":"ABot-World-0: Tworzenie nieograniczonego interaktywnego świata przy użyciu pojedynczej karty GPU klasy desktop","summary":"ABot-World-0 to model świata wideo sterowany akcjami, który może działać w nieskończonym generowaniu interaktywnego świata na pojedynczej karcie NVIDIA RTX 5090 GPU w rozdzielczości 720P, z maksymalnie 16 FPS, opóźnieniem akcji do pierwszej klatki wynoszącym 1,2 sekundy oraz szczytowym budżetem pamięci wideo około 19 GiB. Model wykorzystuje jednolity zsynchronizowany interfejs klawiatury do akcji klatkowych, rozwiązuje problem dryfu autoregresji na długich odcinkach dzięki etapowi dopasowania dystrybucji LongForcing oraz integruje lekkie dekodery VAE, wnioskowanie niskobitowe DiT i buforowanie lokalnego kontekstu klucz-wartość KV jako optymalizacje systemowe.","category":"论文研究","source":"HuggingFace Daily Papers（社区热门论文）","aggregationSource":"HuggingFace Daily Papers（社区热门论文）","pageTitle":"ABot-World-0: Tworzenie nieograniczonego interaktywnego świata przy użyciu pojedynczej karty GPU klasy desktop - Aioga Wiadomości AI","description":"ABot-World-0 to model świata wideo sterowany akcjami, który może działać w nieskończonym generowaniu interaktywnego świata na pojedynczej karcie NVIDIA RTX 5090 GPU w rozdzielczośc...","url":"https://www.aioga.com/pl/news/cmrvls9qg03jtbihbfhtx203y/","contentTranslated":true,"sourceHash":"fe0325e6e2c59a89","translatedAt":"2026-07-22T16:52:57.178Z"}}}}