i2rt_robotics's profile picture. Open-source, robust robotic hardware that you can actually deploy in real world.  Email support@i2rt.com for inquiries!

i2RT Robotics

@i2rt_robotics

Open-source, robust robotic hardware that you can actually deploy in real world. Email [email protected] for inquiries!

置顶

Mama I’m on TV 🥹

i2rt_robotics's tweet image. Mama I’m on TV 🥹

Excited to announce that we have raised $120M in our Series A to advance the frontier of general-purpose high-performance robots. 🤖 The new funding will accelerate progress towards our mission of bringing foundation-model powered robots to everyone, everywhere. Read more 👇



I could have helped but you chose to let me sit there and watch 🧐

Infra upgrade day



i2RT Robotics 已转帖

On-policy probing from a strong base model + off-policy residual learning + distillation might be a right formula for VLA post-training. The most interesting part to me is the intriguing cross-task generalization after distillation. Check out the PLD project led by @_wenlixiao

What if robots could improve themselves by learning from their own failures in the real-world? Introducing 𝗣𝗟𝗗 (𝗣𝗿𝗼𝗯𝗲, 𝗟𝗲𝗮𝗿𝗻, 𝗗𝗶𝘀𝘁𝗶𝗹𝗹) — a recipe that enables Vision-Language-Action (VLA) models to self-improve for high-precision manipulation tasks. PLD…



it’s just matter of time before we see YAMs start assembling YAMs 🥹

What if robots could improve themselves by learning from their own failures in the real-world? Introducing 𝗣𝗟𝗗 (𝗣𝗿𝗼𝗯𝗲, 𝗟𝗲𝗮𝗿𝗻, 𝗗𝗶𝘀𝘁𝗶𝗹𝗹) — a recipe that enables Vision-Language-Action (VLA) models to self-improve for high-precision manipulation tasks. PLD…



Behind the scene: how are YAMs made? You’re watching the manufacturing process of J2 actuator on YAM Pro and YAM Ultra.


Who wants to see what’s inside a YAM that makes it so delicious? when i say vertical integration i mean we make the motor that make the YAM arm and therefore we’re able to achieve the best-in-class system-level performance. Follow and I’ll share more on what’s inside the YAM…

i2rt_robotics's tweet image. Who wants to see what’s inside a YAM that makes it so delicious?

when i say vertical integration i mean we make the motor that make the YAM arm and therefore we’re able to achieve the best-in-class system-level performance.

Follow and I’ll share more on what’s inside the YAM…
i2rt_robotics's tweet image. Who wants to see what’s inside a YAM that makes it so delicious?

when i say vertical integration i mean we make the motor that make the YAM arm and therefore we’re able to achieve the best-in-class system-level performance.

Follow and I’ll share more on what’s inside the YAM…

Thanks @DJiafei !! One of the missions here at i2RT is to democratize capable hardware for physical intelligence 🚀

This is why I’m bullish that the coming years will mark a turning point for robotics: as capable hardware becomes more accessible and affordable, large-scale adoption and real-world impact will finally be within reach. Great arms from @i2rt_robotics !

DJiafei's tweet image. This is why I’m bullish that the coming years will mark a turning point for robotics: as capable hardware becomes more accessible and affordable, large-scale adoption and real-world impact will finally be within reach. 

Great arms from @i2rt_robotics !


Dancing in the moonlight 👯‍♂️ YAM EOL testing. All YAM customers, send me your vin number etched at bottom and I’ll share your arm’s little cute dance video.


Each increment here is 2 µm, 1/35 of the thickness of your hair Repeatability matters. Even for low cost arms. (Please watch because someone spent 1 hour setting this up today)


That’s why we grind our own gears, wind our own motor and fabricate controller board all in-house.

True system-level performance comes from controlling every layer, not just assembling off-the-shelf parts.



This is what fundamentally makes i2RT different - vertical integration from open-source code down to the motor/gearbox/ controller manufacturing. There are no shortcuts to system-level high performance.

i2rt_robotics's tweet image. This is what fundamentally makes i2RT different - vertical integration from open-source code down to the motor/gearbox/ controller manufacturing. 

There are no shortcuts to system-level high performance.

Hold my beer 🤡

This dude shows you how to fold a t shirt in 3 seconds



Bring them YAMs out 🦾

Leader arms are mounted on the base. Teleop setup next! #robotics

saba_khalilnaji's tweet image. Leader arms are mounted on the base. Teleop setup next!

#robotics


Turns out I’m just really good at t-shirt folding …

🚀 Introducing SARM: Stage-Aware Reward Modeling for Long-Horizon Robot Manipulation Robots struggle with tasks like folding a crumpled T-shirt—long, contact-rich, and hard to label. We propose a scalable reward modeling framework to fix that. 1/n



We’re probably one of the few hardware* robotics companies out there that has literally no hype on the internet but large fleet in commercial deployment (by our customers). Be grounded. Serve your customers. Hardware and scaled manufacturing are dirty but someone has to move the…


If you’re serious about deploying useful robot in real world this is the best the form factor on the market at the moment.

a 6dof robotic arm isn’t just a machine; it’s the closest thing we have to replicating the human arm in robotics. why 6 dof? 3 dof cover position (x, y, z) 3 dof cover orientation (roll, pitch, yaw) together → you can place the end-effector anywhere, at any angle. what makes…

oprydai's tweet image. a 6dof robotic arm isn’t just a machine; it’s the closest thing we have to replicating the human arm in robotics.

why 6 dof?

3 dof cover position (x, y, z)
3 dof cover orientation (roll, pitch, yaw)
together → you can place the end-effector anywhere, at any angle.

what makes…


“Shirts are not gonna fold themselves” - everyone’s mom, 2024

We are really impressed with @DynaRobotics robot folding demo at @foxglove's Actuate conference. We've come a long way since the early PR2 folding research! youtu.be/caVo6EJFV8s?fe…



United States 趋势

Loading...

Something went wrong.


Something went wrong.