Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

Introducing Axis Dataset V1 - the simulation dataset for scalable robot manipulation. Can noisy, crowdsourced simulation data support embodied pretraining? Yes—with enough coverage and diversity. Continual pretraining on AXIS dataset V1 lifts π0.5 from 83.9% to 88.8% on LIBERO-Plus, with performance improving consistently as the pretraining data scales from...

114,840 Aufrufe • vor 3 Monaten •via X (Twitter)

34 Kommentare

Profilbild von Stone Tao
Stone Taovor 3 Monaten

don’t want to put too much shade on the progress (i think sim datasets can be good ideas if executed well!), but improving on libero (which is also a sim benchmark) means very little unfortunately

Profilbild von Thor
Thorvor 3 Monaten

Let's goo 🔥

Profilbild von Hazzy
Hazzyvor 3 Monaten

you edit too much intern

Profilbild von KABILA
KABILAvor 2 Monaten

Cookingg

Profilbild von Jontu
Jontuvor 2 Monaten

This is a strong proof that diverse simulation data can meaningfully improve embodied AI.

Profilbild von boyBrands_
boyBrands_vor 3 Monaten

He has arrived🔥

Profilbild von Sami
Samivor 3 Monaten

So much excited for this

Profilbild von Apurbo
Apurbovor 2 Monaten

We are so excited about this 🤘

Profilbild von AL Shifat
AL Shifatvor 3 Monaten

Axis going to make history

Profilbild von dviet.eth
dviet.ethvor 3 Monaten

This is exactly the kind of infrastructure Physical AI needs. Quality data compounds, and so does robot intelligence. Excited to see Axis Dataset V1 keep growing 😆

Profilbild von Noodle
Noodlevor 3 Monaten

cool to see those teleop sessions actually pushing the libero-plus numbers up

Profilbild von Acee (✱,✱)
Acee (✱,✱)vor 3 Monaten

Keep cooking 👌

Profilbild von Piggie 🧡|Bird
Piggie 🧡|Birdvor 3 Monaten

Overall success rate increased by 5.8%.

Profilbild von 0xmowra.eth
0xmowra.ethvor 3 Monaten

Axis V1 dataset boosting robot learning with crowdsourced simulation data impressive

Profilbild von Red
Redvor 2 Monaten

Great work

Profilbild von | MinhDucc | Shark
| MinhDucc | Sharkvor 3 Monaten

Axis LFG 😍

Profilbild von sasafonda✱,✱)
sasafonda✱,✱)vor 3 Monaten

gaxis

Profilbild von namnew
namnewvor 3 Monaten

gAxis

Profilbild von VINAY (✱,✱)
VINAY (✱,✱)vor 3 Monaten

Axis cooking 🔥

Profilbild von savagezz
savagezzvor 3 Monaten

cooking something

Profilbild von Hoangdong (✱,✱)
Hoangdong (✱,✱)vor 3 Monaten

gAxis

Profilbild von clover
clovervor 3 Monaten

move gaxis

Profilbild von vinhle (✱,✱) 🍚 ⛓ ./
vinhle (✱,✱) 🍚 ⛓ ./vor 3 Monaten

Exciting to see the potential of crowdsourced simulation data in robotics, and impressive results from continual pretraining on Axis Dataset V1.

Profilbild von Aurora Lux
Aurora Luxvor 3 Monaten

AXIS Dataset V1 is a strong step forward for embodied AI—scalable, diverse simulation data is proving to be a powerful foundation for better robot learning. 🤖🚀

Profilbild von cbeta.base.eth
cbeta.base.ethvor 3 Monaten

lets goo team fight fight fight

Profilbild von rara
raravor 3 Monaten

gaxis lfg

Profilbild von Linh (✱,✱)
Linh (✱,✱)vor 3 Monaten

Axis Dataset V1 is impressive

Profilbild von Nova
Novavor 3 Monaten

The bot is terrible. Please update it. Its performance is like a primitive human living in a cave 😤

Profilbild von MAS_Enpici (✱,✱)
MAS_Enpici (✱,✱)vor 3 Monaten

big update and new insight

Profilbild von yellowopus
yellowopusvor 3 Monaten

lets cooking bro

Profilbild von KADAFER
KADAFERvor 3 Monaten

let him cook 🔥

Profilbild von dondonat (✱,✱)
dondonat (✱,✱)vor 3 Monaten

gaxis keep building

Profilbild von sagar
sagarvor 3 Monaten

Let's goo 🙌

Profilbild von Gujil Ruipa
Gujil Ruipavor 3 Monaten

Open datasets like this can accelerate robotics research through collaborative innovation

Ähnliche Videos

In my past research experience, finding or developing an appropriate simulation environment, dataset, and benchmark has always been a challenge. Missing features, limited support, or unexpected bugs often occupied my days and nights. Moreover, current simulation platforms are relatively fragmented—making it challenging to replicate the success of the RT-X dataset in unifying community efforts. Introducing RoboVerse, we provide a unified platform, dataset, and benchmark for scalable and generalizable robot learning. We hope to build a shared foundation to combine the community efforts. RoboVerse includes: MetaSim: We carefully designed a configuration system and a universal interface to align current robotic simulators. With MetaSim, you can use any simulator with the same code—bringing together the community’s diverse efforts under one framework! RoboVerse Dataset and Benchmark: We unify popular simulation environments and benchmarks into a single cohesive system and introduce the RoboVerse dataset—a large-scale, high-quality synthetic dataset. Additionally, we propose a standardized benchmark across both imitation learning and reinforcement learning. A cool feature enabled by our unified framework: Hybrid Simulation! You can now integrate physics engines and renderers from different simulators—e.g., using MuJoCo precise physics with Isaac photorealistic rendering. This not only elevates simulation fidelity but also significantly enhances real-world transfer performance across complex robotic applications. Hopefully, our team’s efforts could serve the robotic community to thrive vibrantly in the years to come. RoboVerse is open-sourced🥳!!! Project Page: Documentation: Github Repo: Paper:

Haoran Geng

84,425 Aufrufe • vor 1 Jahr