正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

A video world model can make a convincing clip and still get the physics wrong. Our researchers just released Physis-Lang, an open self-evolving framework that adds physics reasoning to video captions. The captions explain why and how a scene unfolds. We use them to fine-tune world models and add... show more

NVIDIA AI

347,413 subscribers

31,266 次观看 • 14 天前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

10 条评论

NVIDIA AI 的头像

NVIDIA AI14 天前

Physis-Lang with NVIDIA Cosmos 3 Super and Nano rank #1 and #2 on the Physics-IQ Verified image-to-video leaderboard. The benchmark tests how well generated videos follow real-world physics. Full rankings:

Xianzheng Ma 的头像

Xianzheng Ma14 天前

Happy to have contributed to Physis-Lang as a co-first author! Such a rewarding collaboration with the team. Excited to see our work shared here, and grateful to everyone who helped make it happen!

NVIDIA Robotics 的头像

NVIDIA Robotics14 天前

👏👏👏

The Lucky Lighthouse 的头像

The Lucky Lighthouse14 天前

Wanna see the solved equation?

Marie AI 的头像

Marie AI14 天前

Amezing

Emir Çitak 的头像

Emir Çitak14 天前

Explaining why the scene happens before drawing it is a smart order. Most weird AI videos look fine frame by frame and break exactly at the why.

EKOS _ AGI 🦊 🇮🇷 的头像

EKOS _ AGI 🦊 🇮🇷14 天前

You can test and usu EKOS 🎉🔥

Ritik Bompilwar 的头像

Ritik Bompilwar14 天前

🔥🔥

Puneet 的头像

Puneet14 天前

$NVDA next 15 years are looking very bad 👀🧐😳

Tal 的头像

Tal14 天前

@grok what did they do here? Does a language model enrich the prompt to make the video model to produce more physics based output?

相关视频

🕹️We are excited to introduce "ChronoEdit: Towards Temporal Reasoning for Image Editing and World Simulation" ChronoEdit reframes image editing as a video generation task to encourage temporal consistency. It leverages a temporal reasoning stage that denoises with “video reasoning tokens” to "reason" on physically plausible edits. See the attached video for results. Project Page: Arxiv: Code and model are coming.

🕹️We are excited to introduce "ChronoEdit: Towards Temporal Reasoning for Image Editing and World Simulation" ChronoEdit reframes image editing as a video generation task to encourage temporal consistency. It leverages a temporal reasoning stage that denoises with “video reasoning tokens” to "reason" on physically plausible edits. See the attached video for results. Project Page: Arxiv: Code and model are coming.

Huan Ling

37,048 次观看 • 1 年前

LLM-grounded Diffusion: Enhancing Prompt Understanding of Text-to-Image Diffusion Models with Large Language Models paper page: github: Recent advancements in text-to-image generation with diffusion models have yielded remarkable results synthesizing highly realistic and diverse images. However, these models still encounter difficulties when generating images from prompts that demand spatial or common sense reasoning. We propose to equip diffusion models with enhanced reasoning capabilities by using off-the-shelf pretrained large language models (LLMs) in a novel two-stage generation process. First, we adapt an LLM to be a text-guided layout generator through in-context learning. When provided with an image prompt, an LLM outputs a scene layout in the form of bounding boxes along with corresponding individual descriptions. Second, we steer a diffusion model with a novel controller to generate images conditioned on the layout. Both stages utilize frozen pretrained models without any LLM or diffusion model parameter optimization. We validate the superiority of our design by demonstrating its ability to outperform the base diffusion model in accurately generating images according to prompts that necessitate both language and spatial reasoning. Additionally, our method naturally allows dialog-based scene specification and is able to handle prompts in a language that is not well-supported by the underlying diffusion model.

LLM-grounded Diffusion: Enhancing Prompt Understanding of Text-to-Image Diffusion Models with Large Language Models paper page: github: Recent advancements in text-to-image generation with diffusion models have yielded remarkable results synthesizing highly realistic and diverse images. However, these models still encounter difficulties when generating images from prompts that demand spatial or common sense reasoning. We propose to equip diffusion models with enhanced reasoning capabilities by using off-the-shelf pretrained large language models (LLMs) in a novel two-stage generation process. First, we adapt an LLM to be a text-guided layout generator through in-context learning. When provided with an image prompt, an LLM outputs a scene layout in the form of bounding boxes along with corresponding individual descriptions. Second, we steer a diffusion model with a novel controller to generate images conditioned on the layout. Both stages utilize frozen pretrained models without any LLM or diffusion model parameter optimization. We validate the superiority of our design by demonstrating its ability to outperform the base diffusion model in accurately generating images according to prompts that necessitate both language and spatial reasoning. Additionally, our method naturally allows dialog-based scene specification and is able to handle prompts in a language that is not well-supported by the underlying diffusion model.

AK

83,681 次观看 • 3 年前

Okay, the part of the Odyssey-3 launch I keep thinking about is the physics. Their framing is simple: physics is to world models what coding is to language models. If a model can predict how things move and react, robots and AI agents can work out what happens next. Odyssey-3 Pro posted the highest reported Physics-IQ Verified video-to-video score. That's 66.1 with best-of-8 sampling. The research preview is open if you want to try it yourself:

Okay, the part of the Odyssey-3 launch I keep thinking about is the physics. Their framing is simple: physics is to world models what coding is to language models. If a model can predict how things move and react, robots and AI agents can work out what happens next. Odyssey-3 Pro posted the highest reported Physics-IQ Verified video-to-video score. That's 66.1 with best-of-8 sampling. The research preview is open if you want to try it yourself:

Chidanand Tripathi

107,516 次观看 • 6 天前

Agentic Real2Sim is an interesting take on world modeling. Instead of learning physics directly from videos, it lets a Vision-Language Agent reconstruct a physics simulator from a single real-world video. The result is an editable, executable world model that can be queried, simulated, and improved.

Agentic Real2Sim is an interesting take on world modeling. Instead of learning physics directly from videos, it lets a Vision-Language Agent reconstruct a physics simulator from a single real-world video. The result is an editable, executable world model that can be queried, simulated, and improved.

Robots Digest 🤖

12,075 次观看 • 2 个月前

You might be wondering... How does Project Genie work?🤔 Great question. Project Genie is a prototype web app powered by several of our most advanced AI models, each bringing a unique capability to the equation. From Genie 3's ability to simulate the physics and interactions of any scenario, to Nano Banana's sketching and design, to Gemini's advanced world knowledge and reasoning, this combination results in Project Genie's unique, interactive, (and really cool) user experience. But don't just take our word for it, try it today at:

You might be wondering... How does Project Genie work?🤔 Great question. Project Genie is a prototype web app powered by several of our most advanced AI models, each bringing a unique capability to the equation. From Genie 3's ability to simulate the physics and interactions of any scenario, to Nano Banana's sketching and design, to Gemini's advanced world knowledge and reasoning, this combination results in Project Genie's unique, interactive, (and really cool) user experience. But don't just take our word for it, try it today at:

Google AI

136,762 次观看 • 8 个月前

NVIDIA released a video showing their new Alpamayo In-Car Reasoning Model. An AI that explains decisions, responds to passengers, and drives. How it works: • Verbalized Reasoning: In complex situations, the car selectively explains its reasoning aloud, giving passengers insight into the Al's decisions and helping to build trust. • Natural Language Q&A: Riders can ask questions about the car's behavior or the surrounding environment, and Alpamayo responds in real time. • Language Instructions and Commands: Passengers can request actions such as changing lanes, speeding up, taking a detour, or pulling over, and Alpamayo performs them naturally.

NVIDIA released a video showing their new Alpamayo In-Car Reasoning Model. An AI that explains decisions, responds to passengers, and drives. How it works: • Verbalized Reasoning: In complex situations, the car selectively explains its reasoning aloud, giving passengers insight into the Al's decisions and helping to build trust. • Natural Language Q&A: Riders can ask questions about the car's behavior or the surrounding environment, and Alpamayo responds in real time. • Language Instructions and Commands: Passengers can request actions such as changing lanes, speeding up, taking a detour, or pulling over, and Alpamayo performs them naturally.

Nic Cruz Patane

87,841 次观看 • 6 个月前

Nvidia has just announced Alpamayo 2 Super, an open 34 billion parameter reasoning vision-language-action model designed to accelerate the development of autonomous vehicles. This new model combines the NVIDIA Cosmos 3 Super reasoning model with a 2 billion parameter diffusion-based action expert model, and is post trained with reinforcement learning. The model can return multiple outputs: future trajectory plans, reasoning traces, grounded answers to questions about the scenes, and auto label generation. The model weights are now available for anyone to download on Hugging Face, and the inference code has been posted to GitHub. Distilled models can be deployed commercially without any further permission from Nvidia, and model outputs carry no license conditions. Automakers can distill down a compact version of this model that can run on the Nvidia computer in the car. Major kudos to Nvidia and Jensen Huang for advancing the state of the industry by releasing this as an open model with permissive licensing. Jensen isn't just paying lip service to the idea of open models, Nvidia is actually contributing to the ecosystem — and it's great for their business, because it helps sell more Thor computers that go in the car. Anyone can go download the model and play with it. If you do, let me know what you think. Personally I think it's so cool that we have open weights models that are this advanced, for anyone to download.

Nvidia has just announced Alpamayo 2 Super, an open 34 billion parameter reasoning vision-language-action model designed to accelerate the development of autonomous vehicles. This new model combines the NVIDIA Cosmos 3 Super reasoning model with a 2 billion parameter diffusion-based action expert model, and is post trained with reinforcement learning. The model can return multiple outputs: future trajectory plans, reasoning traces, grounded answers to questions about the scenes, and auto label generation. The model weights are now available for anyone to download on Hugging Face, and the inference code has been posted to GitHub. Distilled models can be deployed commercially without any further permission from Nvidia, and model outputs carry no license conditions. Automakers can distill down a compact version of this model that can run on the Nvidia computer in the car. Major kudos to Nvidia and Jensen Huang for advancing the state of the industry by releasing this as an open model with permissive licensing. Jensen isn't just paying lip service to the idea of open models, Nvidia is actually contributing to the ecosystem — and it's great for their business, because it helps sell more Thor computers that go in the car. Anyone can go download the model and play with it. If you do, let me know what you think. Personally I think it's so cool that we have open weights models that are this advanced, for anyone to download.

Whole Mars Catalog

45,595 次观看 • 2 个月前

This is THE moment of Physical AI! We are officially announcing Cosmos 3: Omnimodal World Models for Physical AI 🚀 - Cosmos 3 is an omnimodal world model: within a unified architecture, it can understand and generate language, images, video, audio, and actions. - It is not just a VLM, not just a video generator, not just an audio-visual generative model, and not just a physics simulator / world-action model. It can understand images and videos, generate images, videos, and audio, simulate future worlds, predict actions, and generate robot policies—enabling models to truly begin to “touch the world.” - Cosmos 3 is the #1 open-weight reasoner / T2I / I2V / robot policy across many benchmarks. Huge thanks to every teammate who fought side by side on this journey—from architecture, data, training, infra, serving, and evaluation to post-training. Every part of this project carries an incredible amount of hard work. This was my first time leading a project as Tech Lead, and I feel truly fortunate. The future of Physical AI needs models that can not only “see” and “describe” the world, but also “imagine,” “simulate,” and “act”—and eventually close the loop with the real world. I hope Cosmos 3 can become an important starting point for this direction, and I’m excited to push Physical AI into its next stage together with the open-source community. Welcome to the era of Physical AI. HuggingFace: Project Website: Code:

This is THE moment of Physical AI! We are officially announcing Cosmos 3: Omnimodal World Models for Physical AI 🚀 - Cosmos 3 is an omnimodal world model: within a unified architecture, it can understand and generate language, images, video, audio, and actions. - It is not just a VLM, not just a video generator, not just an audio-visual generative model, and not just a physics simulator / world-action model. It can understand images and videos, generate images, videos, and audio, simulate future worlds, predict actions, and generate robot policies—enabling models to truly begin to “touch the world.” - Cosmos 3 is the #1 open-weight reasoner / T2I / I2V / robot policy across many benchmarks. Huge thanks to every teammate who fought side by side on this journey—from architecture, data, training, infra, serving, and evaluation to post-training. Every part of this project carries an incredible amount of hard work. This was my first time leading a project as Tech Lead, and I feel truly fortunate. The future of Physical AI needs models that can not only “see” and “describe” the world, but also “imagine,” “simulate,” and “act”—and eventually close the loop with the real world. I hope Cosmos 3 can become an important starting point for this direction, and I’m excited to push Physical AI into its next stage together with the open-source community. Welcome to the era of Physical AI. HuggingFace: Project Website: Code:

Max Zhaoshuo Li 李赵硕

1,080,311 次观看 • 4 个月前

Elon Musk on first principles. "A good framework for thinking is physics… first principles reasoning." "Boil things down to their fundamental truths and reason up from there, as opposed to reasoning by analogy." Most of life runs on analogy — copy what others do with slight variations, because that's how you get through the day. New work can't. If the answer isn't already on the shelf, you have to drop to physics: what's actually true, then build up. • Analogy = copy + tweak • First principles = fundamental truths, then reason up • Something new → physics approach "But when you want to do something new, you have to apply the physics approach."

Elon Musk on first principles. "A good framework for thinking is physics… first principles reasoning." "Boil things down to their fundamental truths and reason up from there, as opposed to reasoning by analogy." Most of life runs on analogy — copy what others do with slight variations, because that's how you get through the day. New work can't. If the answer isn't already on the shelf, you have to drop to physics: what's actually true, then build up. • Analogy = copy + tweak • First principles = fundamental truths, then reason up • Something new → physics approach "But when you want to do something new, you have to apply the physics approach."

Tesla Owners Silicon Valley

31,656 次观看 • 1 个月前

Let’s imagine: we have a girl — from thinking and reasoning, to natural voice dialogue, all the way to expressive, physics-perfect motion… all handled by AI. Wouldn't that be AMAZING?? AI WAIFU IS ABOUT TO COME TRUE!!!! 😍

Let’s imagine: we have a girl — from thinking and reasoning, to natural voice dialogue, all the way to expressive, physics-perfect motion… all handled by AI. Wouldn't that be AMAZING?? AI WAIFU IS ABOUT TO COME TRUE!!!! 😍

GiantessPlayground

20,110 次观看 • 6 个月前

🚨 Elon Musk on Grok 3.5: "With Grok 3.5, that is about to be released, it's trying to reason from first principles. The focus of Grok 3.5 is to find the fundamentals of physics and applying physics tools across all lines of reasoning and to aspire to truth with minimal error. So apply kind of the tools of physics to thinking, and you know physics, if you violate conservation of energy or momentum, then you're either going to get a Nobel Prize, or you're wrong. We aim to get to truth with acknowledged error, but minimize that error over time. And I think that's actually extremely important for AI safety. So I've put a lot for a long time about AI safety, and conclusion is the old maxim that honesty is the best policy."

🚨 Elon Musk on Grok 3.5: "With Grok 3.5, that is about to be released, it's trying to reason from first principles. The focus of Grok 3.5 is to find the fundamentals of physics and applying physics tools across all lines of reasoning and to aspire to truth with minimal error. So apply kind of the tools of physics to thinking, and you know physics, if you violate conservation of energy or momentum, then you're either going to get a Nobel Prize, or you're wrong. We aim to get to truth with acknowledged error, but minimize that error over time. And I think that's actually extremely important for AI safety. So I've put a lot for a long time about AI safety, and conclusion is the old maxim that honesty is the best policy."

DogeDesigner

199,274 次观看 • 1 年前

🚨BREAKING: Emperor Kekius Maximus just revealed insights on Grok 3.5 which is dropping soon: > Grok 3.5 will reason from physics based first principles > To find the fundamentals of physics and apply physics tools across all lines of reasoning > Goal is to be truthful with minimal error > We have and will make mistakes but we aspire to correct them very quickly Message for developer community: “Tell us what you need, where we are wrong and how can we make Grok better and xAI will make it happen” —Elon Musk ACCELERATE

🚨BREAKING: Emperor Kekius Maximus just revealed insights on Grok 3.5 which is dropping soon: > Grok 3.5 will reason from physics based first principles > To find the fundamentals of physics and apply physics tools across all lines of reasoning > Goal is to be truthful with minimal error > We have and will make mistakes but we aspire to correct them very quickly Message for developer community: “Tell us what you need, where we are wrong and how can we make Grok better and xAI will make it happen” —Elon Musk ACCELERATE

NIK

210,890 次观看 • 1 年前

#PixVerseWorldModel is doing something pretty interesting with AI video. One of the things that stood out to me is the WASD experience. AI video usually follows a simple workflow: write a prompt, generate a clip, and watch the result. With PixVerse World Model, you can actually move around inside the generated world, control the camera, and use prompts to influence the environment, characters, and actions as you move through the scene. The interesting part is how movement and prompting work together. Instead of generating a completely new clip every time you want something to change, you can keep interacting with the world and see how it responds. Text prompts start to feel less like instructions for generating a video and more like another way of interacting with the world itself. It makes PixVerse World Model feel less like traditional AI video and more like a playable, generated environment. Maybe the more interesting question for AI video isn’t just “what can it generate?” It’s “what can we actually do once we’re inside it?” Try now:

#PixVerseWorldModel is doing something pretty interesting with AI video. One of the things that stood out to me is the WASD experience. AI video usually follows a simple workflow: write a prompt, generate a clip, and watch the result. With PixVerse World Model, you can actually move around inside the generated world, control the camera, and use prompts to influence the environment, characters, and actions as you move through the scene. The interesting part is how movement and prompting work together. Instead of generating a completely new clip every time you want something to change, you can keep interacting with the world and see how it responds. Text prompts start to feel less like instructions for generating a video and more like another way of interacting with the world itself. It makes PixVerse World Model feel less like traditional AI video and more like a playable, generated environment. Maybe the more interesting question for AI video isn’t just “what can it generate?” It’s “what can we actually do once we’re inside it?” Try now:

Nayeem Sheikh

14,959 次观看 • 21 天前

The next frontier of autonomous driving is unlocked by reasoning models. NVIDIA Alpamayo brings together open AI models with reasoning capabilities, closed-loop simulation tools, and massive real-world driving datasets. Alpamayo 1 is a vision–language–action model that explains its own decisions through explicit reasoning traces, enabling trustworthy, humanlike decision-making. Together with NVIDIA’s Physical AI dataset and AlpaSim simulation, Alpamayo provides the tools and scale required to enable level 4 autonomous vehicles. ▶️ Watch now:

The next frontier of autonomous driving is unlocked by reasoning models. NVIDIA Alpamayo brings together open AI models with reasoning capabilities, closed-loop simulation tools, and massive real-world driving datasets. Alpamayo 1 is a vision–language–action model that explains its own decisions through explicit reasoning traces, enabling trustworthy, humanlike decision-making. Together with NVIDIA’s Physical AI dataset and AlpaSim simulation, Alpamayo provides the tools and scale required to enable level 4 autonomous vehicles. ▶️ Watch now:

NVIDIA DRIVE

35,324 次观看 • 9 个月前

this is f**king dangerous someone just figured out how to get fable 5 reasoning in opus 4.8 with one prompt you have less than 24 hours to set this up as this takes away 20% of your usage limit here is how: 1) Paste the prompt and ask it to "make an operating manual" 2) Create a New Claude Project and "Add this .md in Project instructions 3) Use Opus 4.8 High/Max and now you have Fable 5 reasoning at 1/5th the price save and bookmark it no matter what

this is f**king dangerous someone just figured out how to get fable 5 reasoning in opus 4.8 with one prompt you have less than 24 hours to set this up as this takes away 20% of your usage limit here is how: 1) Paste the prompt and ask it to "make an operating manual" 2) Create a New Claude Project and "Add this .md in Project instructions 3) Use Opus 4.8 High/Max and now you have Fable 5 reasoning at 1/5th the price save and bookmark it no matter what

Avid

954,976 次观看 • 3 个月前

NVIDIA just introduced Cosmos, a platform for world foundation models designed for robotics. ⦿ It features advanced tokenizers, an AI-accelerated data pipeline, and integration with NVIDIA Omniverse. Humanoid makers 1X, Figure, and Agility are among the first to adopt Cosmos. ⦿ Cosmos generates synthetic, physics-based data, accelerating model training and customization. ⦿ It also features a CUDA-accelerated data processing pipeline that enables developers to process, curate, and label 20 million hours of videos in 14 days using the NVIDIA Blackwell platform.

NVIDIA just introduced Cosmos, a platform for world foundation models designed for robotics. ⦿ It features advanced tokenizers, an AI-accelerated data pipeline, and integration with NVIDIA Omniverse. Humanoid makers 1X, Figure, and Agility are among the first to adopt Cosmos. ⦿ Cosmos generates synthetic, physics-based data, accelerating model training and customization. ⦿ It also features a CUDA-accelerated data processing pipeline that enables developers to process, curate, and label 20 million hours of videos in 14 days using the NVIDIA Blackwell platform.

The Humanoid Hub

129,481 次观看 • 1 年前

A little video that - explains the relationship between conventional LLMs and reasoning models (and agents), - philosophizes a about "from scratch" approaches, - and explains how to install Python & PyTorch requirements with uv.

A little video that - explains the relationship between conventional LLMs and reasoning models (and agents), - philosophizes a about "from scratch" approaches, - and explains how to install Python & PyTorch requirements with uv.

Sebastian Raschka

60,094 次观看 • 1 个月前