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。

We just released the world's first open-source humanoid physical strike AI! The AI is trained with large-scale self-play RL with a pre-trained ASE model from Jason Peng. Thank ZiyiBird for this amazing work.

Flood Sung

2,125 subscribers

101,777 次观看 • 3 年前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

10 条评论

Mike Ma (AI x Finance) 的头像

Mike Ma (AI x Finance)3 年前

@xbpeng4 @ZiyiBird DARPA be like -

NICE 的头像

NICE1 年前

Stay competitive by balancing cutting-edge AI with automation tools. Forrester shows how.

Gavriel State 的头像

Gavriel State3 年前

@xbpeng4 @ZiyiBird This is very cool! Video capture rate is a bit too low though, so it's hard to tell if they are learning to parry properly. It would be great to see how the behaviour changes with early termination based on received damage.

Crystalwizard 的头像

Crystalwizard3 年前

@xbpeng4 @ZiyiBird next step, sending the AIs to war, and then a lottery where people find out they would have died and report to death chambers. oh, wait, there's a science fiction movie/tv_episode/book/story about that ...

DenysM 的头像

DenysM3 年前

@xbpeng4 @ZiyiBird Looks cool!

James 的头像

James3 年前

@xbpeng4 @ZiyiBird I don't see a human model in the assets folder

David Jianhong Wang 的头像

David Jianhong Wang3 年前

@xbpeng4 @ZiyiBird An interesting simulator! Good work!

James 的头像

James3 年前

@xbpeng4 @ZiyiBird It would be so helpful for robotics if you included an env with wheeled robot as physics is rather different. I'm interested in robot learning to navigate in environment with obstacles and people.

Nikolaj Goodger 的头像

Nikolaj Goodger3 年前

@xbpeng4 @ZiyiBird Cool. Imagine a robot trained like this in suit of armour.

inKstall Educare 的头像

inKstall Educare3 年前

@RLCommunity8 @xbpeng4 @ZiyiBird

相关视频

AI is trained on ꧁IP꧂ Story will unlock IP as one of the world's largest asset classes. Jason Zhao chops it up with Frank Chaparro from The Block.

AI is trained on ꧁IP꧂ Story will unlock IP as one of the world's largest asset classes. Jason Zhao chops it up with Frank Chaparro from The Block.

Story

45,644 次观看 • 1 年前

I remember sitting down at my desk for the first time, Quan Vuong showing me the starter project: let’s make pre-trained models work without fine-tuning? With π0.7, our pre-trained model works out of the box across so many tasks, matching or even outperforming SFT or RL specialists!

I remember sitting down at my desk for the first time, Quan Vuong showing me the starter project: let’s make pre-trained models work without fine-tuning? With π0.7, our pre-trained model works out of the box across so many tasks, matching or even outperforming SFT or RL specialists!

Allen Ren

14,637 次观看 • 2 个月前

This is crazy... Skywork AI just dropped SkyReels-V2, the world's first open source AI video model that supports unlimited duration. Create infinite-length cinematic videos, straight from your browser. More examples below:

This is crazy... Skywork AI just dropped SkyReels-V2, the world's first open source AI video model that supports unlimited duration. Create infinite-length cinematic videos, straight from your browser. More examples below:

Angry Tom

302,506 次观看 • 1 年前

German robotics company NEURA Robotics has created NEURA Gym: a large-scale, physical AI gym and training ground where hundreds of robots, including the humanoid 4NE-1, learn through real-world interactions. The company is collaborating with the Neuraverse ecosystem to create the world's largest repository of physical training data, promoting the application of humanoid robots.

German robotics company NEURA Robotics has created NEURA Gym: a large-scale, physical AI gym and training ground where hundreds of robots, including the humanoid 4NE-1, learn through real-world interactions. The company is collaborating with the Neuraverse ecosystem to create the world's largest repository of physical training data, promoting the application of humanoid robots.

CyberRobo

71,698 次观看 • 9 个月前

Introducing `AutoRL` 📈 The world's simplest way to train a task-specific LLM with RL. *Just write a SENTENCE describing the model you want.* A chain of AI systems will generate data + rubrics and train a model for you. Powered by ART, it's open source. Link in thread:

Introducing `AutoRL` 📈 The world's simplest way to train a task-specific LLM with RL. Just write a SENTENCE describing the model you want. A chain of AI systems will generate data + rubrics and train a model for you. Powered by ART, it's open source. Link in thread:

Matt Shumer

150,107 次观看 • 11 个月前

We trained a graph-native AI, then let it reason for days, forming a dynamic relational world model on its own - no pre-programming. Emergent hubs, small-world properties, modularity, & scale-free structures arose naturally. The model then exploited compositional reasoning & uncovered uncoded properties from deep synthesis: Materials with memory, microbial repair, self-evolving systems. Video shows it unfolding, made with Grok xAI.

We trained a graph-native AI, then let it reason for days, forming a dynamic relational world model on its own - no pre-programming. Emergent hubs, small-world properties, modularity, & scale-free structures arose naturally. The model then exploited compositional reasoning & uncovered uncoded properties from deep synthesis: Materials with memory, microbial repair, self-evolving systems. Video shows it unfolding, made with Grok xAI.

Markus J. Buehler

359,403 次观看 • 1 年前

⚙️ 404—GEN has released the WORLD'S LARGEST open-source 3D model dataset ⚙️ • 21.5M+ AI-generated 3d models • Open-source and attribution-ready • Larger than all existing 3D datasets combined This is an unmatched scale for Gaussian Splatting research. More details below 🧵 ⬇️

⚙️ 404—GEN has released the WORLD'S LARGEST open-source 3D model dataset ⚙️ • 21.5M+ AI-generated 3d models • Open-source and attribution-ready • Larger than all existing 3D datasets combined This is an unmatched scale for Gaussian Splatting research. More details below 🧵 ⬇️

404

44,735 次观看 • 1 年前

RobotMDM, by Disney Research, combines diffusion-based motion generation with RL to produce physics-aware humanoid motions from text prompts. Trained on human motion data with a reward surrogate for physical feasibility, it ensures realistic motions.

RobotMDM, by Disney Research, combines diffusion-based motion generation with RL to produce physics-aware humanoid motions from text prompts. Trained on human motion data with a reward surrogate for physical feasibility, it ensures realistic motions.

The Humanoid Hub

22,943 次观看 • 1 年前

We just solved creating infinitely long AI videos from a single prompt. We built Latted as the unified workspace for AI video projects. Then, we made the world's first AI generated feature film with it!

We just solved creating infinitely long AI videos from a single prompt. We built Latted as the unified workspace for AI video projects. Then, we made the world's first AI generated feature film with it!

emre

13,930 次观看 • 4 个月前

🔫 Badass! A team at the Mistral AI hackathon in SF trained the 7B open-source model to play DOOM, based on an ASCII representation of the current frame in the game. 🤯 John Carmack

🔫 Badass! A team at the Mistral AI hackathon in SF trained the 7B open-source model to play DOOM, based on an ASCII representation of the current frame in the game. 🤯 John Carmack

👩‍💻 Paige Bailey

857,627 次观看 • 2 年前

At Malaysia Blockchain Week this year, we unveiled world's 1st fully trained AI Agent 🦾 Working with Alliance and Team Secret, we trained an AI-Agent in Counterstrike from Flaxciz’s gameplay. - Training Time: 102 hours - Visual Data: 51 hours of gameplay Watch below

At Malaysia Blockchain Week this year, we unveiled world's 1st fully trained AI Agent 🦾 Working with Alliance and Team Secret, we trained an AI-Agent in Counterstrike from Flaxciz’s gameplay. - Training Time: 102 hours - Visual Data: 51 hours of gameplay Watch below

iAgent

23,812 次观看 • 1 年前

Introducing Paris - world's first decentralized trained open-weight diffusion model. We named it Paris after the city that has always been a refuge for those creating without permission. Paris is open for research and commercial use.

Introducing Paris - world's first decentralized trained open-weight diffusion model. We named it Paris after the city that has always been a refuge for those creating without permission. Paris is open for research and commercial use.

bagel.com

606,561 次观看 • 9 个月前

Weights drop ⚠️ We released our pre-trained model for the cup arrangement task trained on 1400 demos! We aim to enable anyone to deploy UMI on their robot to arrange any "espresso cup with saucer" they buy on Amazon.

Weights drop ⚠️ We released our pre-trained model for the cup arrangement task trained on 1400 demos! We aim to enable anyone to deploy UMI on their robot to arrange any "espresso cup with saucer" they buy on Amazon.

Cheng Chi

25,131 次观看 • 2 年前

World Foundation Models (WFM) are the future of Physical AI, and we believe this future should be open-sourced. We are partnering with Valeo Group to build one of the largest open-source, multi-camera WFMs — redefining how Physical AI interacts with the real world.

World Foundation Models (WFM) are the future of Physical AI, and we believe this future should be open-sourced. We are partnering with Valeo Group to build one of the largest open-source, multi-camera WFMs — redefining how Physical AI interacts with the real world.

NATIX Network

261,023 次观看 • 5 个月前

🤖What if a robot could perform a new task just from a natural language command, with zero demonstrations? Our new work, NovaFlow, makes it possible! We use pre-trained video generative model to create a video of the task, then translate it into a plan for real-world robot execution. 1/6 #Robotics #AI #ZeroShot #Manipulation

🤖What if a robot could perform a new task just from a natural language command, with zero demonstrations? Our new work, NovaFlow, makes it possible! We use pre-trained video generative model to create a video of the task, then translate it into a plan for real-world robot execution. 1/6 #Robotics #AI #ZeroShot #Manipulation

Hongyu Li

105,560 次观看 • 9 个月前

This is amazing research result! Xun Huang and team trained a bidirectional teacher model for motion control, and used Self-Forcing to make it autoregressive. Xun recently gave a great talk at the Real Time Video AI summit about this work, and a few other projects around video world models. Here is the recording.

This is amazing research result! Xun Huang and team trained a bidirectional teacher model for motion control, and used Self-Forcing to make it autoregressive. Xun recently gave a great talk at the Real Time Video AI summit about this work, and a few other projects around video world models. Here is the recording.

Eric is Daydreaming

52,796 次观看 • 8 个月前