Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

End-to-end whole body mobile manipulation on Agility Digit. We ran this demo for ~12 hours at IROS earlier this week video is 1x speed, autonomous, in a previously-unseen location (not in training data).

chris_j_paxton's profile picture

Chris Paxton

37,454 subscribers

11,971 görüntüleme • 3 gün önce •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

11 Yorum

Arjun Subramaniam profil fotoğrafı
Arjun Subramaniam3 gün önce

@agilityrobotics This is very cool!

John Koetsier profil fotoğrafı
John Koetsier3 gün önce

@agilityrobotics looking forward to seeing how much faster, more stable, and more confident Digit 5 is

AAIL profil fotoğrafı
AAIL3 gün önce

@agilityrobotics yooo 12 hours in a room it hasn't seen and it looks so employed

Velkorti profil fotoğrafı
Velkorti3 gün önce

@agilityrobotics 1x speed, autonomous, unseen location. That's the caption every robot video should have.

Humanoid Investing profil fotoğrafı
Humanoid Investing3 gün önce

@agilityrobotics Thanks for sharing Chris!

Hank profil fotoğrafı
Hank2 gün önce

@agilityrobotics but how he know what he should do？instructions？

Mod Ned profil fotoğrafı
Mod Ned2 gün önce

@agilityrobotics Was the shape of the objects also unseen?

のぶぶ profil fotoğrafı
のぶぶ2 gün önce

@agilityrobotics これに何の価値があるというのだろう 専用設計して爆速に動作させなければ商売にならない 単なる玩具 よくわからないものを勝手に適当に処理するよりも即座に停止して管理者を呼ぶ方がマシ

Sumeet Kotasthane profil fotoğrafı
Sumeet Kotasthane3 gün önce

@agilityrobotics The 12 hours tells me more than the clip does, because a short video can be the best take, but a full day in a new place shows how often it fails. How many times did someone have to step in over those 12 hours, and what caused most of them?

:(){ :|:& };: profil fotoğrafı
:(){ :|:& };:2 gün önce

@agilityrobotics Impressive. Also is it 12 hours on one charge or is that 12 hours of swapping batteries in and out?

Mo Better profil fotoğrafı
Mo Better2 gün önce

@agilityrobotics when i watch these videos i think… robotics is so damn weak. ughhhh….

Benzer Videolar

End to end, autonomous whole body motion generation. You can see this demo in “live”(not video) at ICRA2024.
0:27

End to end, autonomous whole body motion generation. You can see this demo in “live”(not video) at ICRA2024.

Tetsuya Ogata / 尾形哲也

42,928 görüntüleme • 2 yıl önce

The most impressive robot demo I saw at IROS was also one of the simplest. Agility had Digit autonomously picking objects from a tote, walking them across the booth, and repeating the task until the tote was empty. I talked with the mastermind behind the demo, Chris Paxton, about how the demo worked: ~500 real-world teleoperated demonstrations for the manipulation policy A whole-body controller trained in simulation Deployed into a completely unseen environment at IROS Roughly 80–90% success in the booth despite crowds, harsh lighting, and plenty of distractions Chris was also very clear that this is still research, not something currently running in production. It was a glimpse at how quickly learned manipulation policies can move from controlled training environments into the messy real world. Lukas Ziegler you might find this convo interesting. #IROS2026 #AgilityRobotics #PhysicalAI #RobotLearning
3:50

The most impressive robot demo I saw at IROS was also one of the simplest. Agility had Digit autonomously picking objects from a tote, walking them across the booth, and repeating the task until the tote was empty. I talked with the mastermind behind the demo, Chris Paxton, about how the demo worked: ~500 real-world teleoperated demonstrations for the manipulation policy A whole-body controller trained in simulation Deployed into a completely unseen environment at IROS Roughly 80–90% success in the booth despite crowds, harsh lighting, and plenty of distractions Chris was also very clear that this is still research, not something currently running in production. It was a glimpse at how quickly learned manipulation policies can move from controlled training environments into the messy real world. Lukas Ziegler you might find this convo interesting. #IROS2026 #AgilityRobotics #PhysicalAI #RobotLearning

Jonathan Stephens

25,591 görüntüleme • 4 gün önce

Excited to share: Figure 01 completing real-world tasks This is end-to-end autonomous We have made advances in our autonomous navigation, learned perception models, manipulation robust to pose variation, & generalizable systems for future applications
1:21

Excited to share: Figure 01 completing real-world tasks This is end-to-end autonomous We have made advances in our autonomous navigation, learned perception models, manipulation robust to pose variation, & generalizable systems for future applications

Brett Adcock

232,949 görüntüleme • 2 yıl önce

The AV industry is hitting a “Data Wall.” Brute-force training on petabytes of data is reaching a dead end. As models improve, further gains get harder. We bypassed this wall with Factored Embodied AI—achieving zero-shot autonomous steering with just 1,000 hours of driving data.
1:17

The AV industry is hitting a “Data Wall.” Brute-force training on petabytes of data is reaching a dead end. As models improve, further gains get harder. We bypassed this wall with Factored Embodied AI—achieving zero-shot autonomous steering with just 1,000 hours of driving data.

Helm.ai

8,650,862 görüntüleme • 9 ay önce

We finished demo on the north end of the @DeltaCenter this week. Crews logged more than 260,000 hours of work in the last 40 days. On our way!
0:32

We finished demo on the north end of the @DeltaCenter this week. Crews logged more than 260,000 hours of work in the last 40 days. On our way!

Ryan Smith

142,191 görüntüleme • 3 ay önce

Tomorrow, 1X is going to release an exciting update on our AI and Autonomy efforts. To give a hint as to what we're about to reveal, here are some of the autonomy videos we've shared in the last 1+ years of scaling up our data engine. Every video in this thread is autonomous, 1X speed
0:25

Tomorrow, 1X is going to release an exciting update on our AI and Autonomy efforts. To give a hint as to what we're about to reveal, here are some of the autonomy videos we've shared in the last 1+ years of scaling up our data engine. Every video in this thread is autonomous, 1X speed

Eric Jang

53,727 görüntüleme • 2 yıl önce

Tracking Anything with Decoupled Video Segmentation paper page: Training data for video segmentation are expensive to annotate. This impedes extensions of end-to-end algorithms to new video segmentation tasks, especially in large-vocabulary settings. To 'track anything' without training on video data for every individual task, we develop a decoupled video segmentation approach (DEVA), composed of task-specific image-level segmentation and class/task-agnostic bi-directional temporal propagation. Due to this design, we only need an image-level model for the target task (which is cheaper to train) and a universal temporal propagation model which is trained once and generalizes across tasks. To effectively combine these two modules, we use bi-directional propagation for (semi-)online fusion of segmentation hypotheses from different frames to generate a coherent segmentation. We show that this decoupled formulation compares favorably to end-to-end approaches in several data-scarce tasks including large-vocabulary video panoptic segmentation, open-world video segmentation, referring video segmentation, and unsupervised video object segmentation.
0:32

Tracking Anything with Decoupled Video Segmentation paper page: Training data for video segmentation are expensive to annotate. This impedes extensions of end-to-end algorithms to new video segmentation tasks, especially in large-vocabulary settings. To 'track anything' without training on video data for every individual task, we develop a decoupled video segmentation approach (DEVA), composed of task-specific image-level segmentation and class/task-agnostic bi-directional temporal propagation. Due to this design, we only need an image-level model for the target task (which is cheaper to train) and a universal temporal propagation model which is trained once and generalizes across tasks. To effectively combine these two modules, we use bi-directional propagation for (semi-)online fusion of segmentation hypotheses from different frames to generate a coherent segmentation. We show that this decoupled formulation compares favorably to end-to-end approaches in several data-scarce tasks including large-vocabulary video panoptic segmentation, open-world video segmentation, referring video segmentation, and unsupervised video object segmentation.

AK

305,683 görüntüleme • 3 yıl önce

we landed on a pretty good workflow for doing parallel work in OpenCode this demo is with git worktrees but i also preview an alternative we're working on at the end this will be in 1.6.0
3:14

we landed on a pretty good workflow for doing parallel work in OpenCode this demo is with git worktrees but i also preview an alternative we're working on at the end this will be in 1.6.0

dax

290,018 görüntüleme • 4 ay önce

Helix is now learning directly from human video data We have already trained on data collected in the real world, including Brookfield residential units To our knowledge, this is the first instance of a humanoid robot learning navigation end-to-end using only human video
0:24

Helix is now learning directly from human video data We have already trained on data collected in the real world, including Brookfield residential units To our knowledge, this is the first instance of a humanoid robot learning navigation end-to-end using only human video

Figure

46,041 görüntüleme • 1 yıl önce

🚨 WATCH: Andy Burnham spoke at dawn earlier today on his pledge to end homelessness in Britain "We ran a scheme every night. 600 beds. But we were the only place doing it. It taught me the only way is to do this together"
1:15

🚨 WATCH: Andy Burnham spoke at dawn earlier today on his pledge to end homelessness in Britain "We ran a scheme every night. 600 beds. But we were the only place doing it. It taught me the only way is to do this together"

Politics UK

651,313 görüntüleme • 2 ay önce

“A huge robot army is loading..” Figure's Helix 02 demo shows end to end "pixels to whole body" autonomy, loading and unloading a dishwasher in a long 4 minute cycle without human intervention !!
0:33

“A huge robot army is loading..” Figure's Helix 02 demo shows end to end "pixels to whole body" autonomy, loading and unloading a dishwasher in a long 4 minute cycle without human intervention !!

SciTech Era

14,195 görüntüleme • 7 ay önce

"I'm not advocating for white supremacy. I am just advocating the end of white silence. The end of white guilt. The end of white shame. Because our culture didn't participate in anything that your culture hadn't participated in previously. We didn't start slavery but we damn sure ended it." This man is on point..listen to the rest of the video.
1:35

"I'm not advocating for white supremacy. I am just advocating the end of white silence. The end of white guilt. The end of white shame. Because our culture didn't participate in anything that your culture hadn't participated in previously. We didn't start slavery but we damn sure ended it." This man is on point..listen to the rest of the video.

Sassafrass84

100,183 görüntüleme • 10 ay önce

GPT-6 Astra is the strongest model for 3D game from a single prompt. We described the game in one sentence > GPT-6 Astra wrote the whole thing - drift physics, combo scoring, near-miss bonuses, speed traps, nitro > out came Street Heat, a full arcade racer running in the browser. The whole pipeline ran end-to-end on Higgsfield Supercomputer.
0:28

GPT-6 Astra is the strongest model for 3D game from a single prompt. We described the game in one sentence > GPT-6 Astra wrote the whole thing - drift physics, combo scoring, near-miss bonuses, speed traps, nitro > out came Street Heat, a full arcade racer running in the browser. The whole pipeline ran end-to-end on Higgsfield Supercomputer.

Higgsfield AI 🧩

175,679 görüntüleme • 1 ay önce

"We talked about it, we moved on from it. At this point in the season, we both realize this is the group we got. So it's definitely not a time for anybody to separate" - Jarred Vanderbilt on his confrontation with JJ Redick earlier this week
0:23

"We talked about it, we moved on from it. At this point in the season, we both realize this is the group we got. So it's definitely not a time for anybody to separate" - Jarred Vanderbilt on his confrontation with JJ Redick earlier this week

Dave McMenamin

157,901 görüntüleme • 5 ay önce

Earlier this week, Democrats and four House Republicans voted to stand up to the President and end this war in Iran. I was proud to take a stance for the American people as we work to rein in this war of choice.
0:56

Earlier this week, Democrats and four House Republicans voted to stand up to the President and end this war in Iran. I was proud to take a stance for the American people as we work to rein in this war of choice.

Rep. Nikki Budzinski

12,486 görüntüleme • 4 ay önce

Training a Whole-Body Control Foundation Model -- new work from my team at Agility A neural network for controlling our humanoid robots which is robust to disturbances, can handle heavy objects, and is a powerful platform for learning new whole-body skills learn more in our blog post ->
1:56

Training a Whole-Body Control Foundation Model -- new work from my team at Agility A neural network for controlling our humanoid robots which is robust to disturbances, can handle heavy objects, and is a powerful platform for learning new whole-body skills learn more in our blog post ->

Chris Paxton

46,193 görüntüleme • 1 yıl önce