Загрузка видео...

Не удалось загрузить видео

На главную

Argus: an open-source robotics data annotation and quality pipeline. With new frontier VLMs like GPT-6 Astra, we can generate rich, high-quality annotations for robotics data for both training and dataset analysis. Argus delivers detailed, timestamp-level annotations while catching issues like mislabeled instructions, sped-up recordings, swapped camera streams, and unflagged...

89,243 просмотров • 1 месяц назад •via X (Twitter)

Комментарии: 33

Фото профиля calix
calix1 месяц назад

Argus takes in any episode in its raw format and outputs the following: - A densely annotated timeline with all robot or gripper actions, state changes, task progression percentages, and more - A record for every operator mistake, along with notes on the timing of the failure, the nature of the failure, the operator’s recovery response, and the ideal fix - Any metadata mistakes - A severity rating for each issue based on the predicted impact on training - Any subgoals and the completion outcome - The goal frame if it is achieved - Overall grade on completion along with a short rationale review

Фото профиля calix
calix1 месяц назад

Argus can also be used on inference and tele-op trajectories captured in production. Here’s an example of a fully annotated trajectory from one of our production facilities.

Фото профиля calix
calix1 месяц назад

Argus is a model-agnostic pipeline and can run on any sufficiently capable VLM. In our test comparisons, Astra and GPT-6.1 Sol produced the densest annotations, which we’ve found is highly correlated with overall accuracy.

Фото профиля calix
calix1 месяц назад

Of the 3,546 episodes we labelled across nine datasets, 27% have at least one issue of medium severity or above, meaning part or all of the episode teaches the model something wrong. We’re releasing the full audit and more than 150K annotations at

Фото профиля calix
calix1 месяц назад

Read the full analysis at and run Argus on your data at For collaborations or questions, contact us at [email protected]. Credit to @ericli_ for undertaking this project.

Фото профиля Henry Lien
Henry Lien1 месяц назад

@PantheonInc Eric 🚀

Фото профиля koko 𝑥𝑠
koko 𝑥𝑠1 месяц назад

@PantheonInc 🚀

Фото профиля andrew kim
andrew kim1 месяц назад

@PantheonInc Incredible

Фото профиля Varun Nair
Varun Nair1 месяц назад

@PantheonInc Pushing the frontier! VLMs are improving so fast, habit. Great harnesses and clarity for training ready data is key to extract signals from raw data and train the right policies!

Фото профиля Mohamed Zidan Cassim
Mohamed Zidan Cassim1 месяц назад

@PantheonInc chef Li @ericli_

Фото профиля carlo agostinelli
carlo agostinelli1 месяц назад

@PantheonInc 🚢

Фото профиля Sambhav Gupta
Sambhav Gupta1 месяц назад

@PantheonInc Eric li rolling the robotics boulder up the hill

Фото профиля Xiao-ke (XK)
Xiao-ke (XK)1 месяц назад

@PantheonInc @ericli_ 🚀🚀🚀

Фото профиля sean jeong
sean jeong1 месяц назад

@PantheonInc 🔥

Фото профиля Rich
Rich1 месяц назад

@PantheonInc Mr Li

Фото профиля Ishan
Ishan1 месяц назад

@PantheonInc So cool

Фото профиля Yang Fan Yun
Yang Fan Yun1 месяц назад

@PantheonInc very very cool stuff @calixo888 @nettedjay 😎

Фото профиля Tian Fang
Tian Fang1 месяц назад

@PantheonInc Very cool.

Фото профиля Nikhil Suresh
Nikhil Suresh1 месяц назад

@PantheonInc 🚢🚢🚢

Фото профиля Felix Huang
Felix Huang1 месяц назад

@PantheonInc Fire contributions for the community

Фото профиля Hudzah
Hudzah1 месяц назад

@PantheonInc beautiful

Фото профиля ChrisPy
ChrisPy1 месяц назад

@PantheonInc 🧑‍🍳

Фото профиля Sufyan Osamah
Sufyan Osamah1 месяц назад

@PantheonInc Swapped camera streams and sped up recordings silently ruin a training run. Good that this is open source.

Фото профиля majjigi
majjigi1 месяц назад

@PantheonInc damn, this is amazing.

Фото профиля Lucas
Lucas1 месяц назад

@PantheonInc Congrats Calix!

Фото профиля Arin
Arin1 месяц назад

@PantheonInc 27% of episodes having training-affecting issues is a wild number for datasets people assumed were clean

Фото профиля joseph
joseph1 месяц назад

@PantheonInc cooking

Фото профиля Federico Martelli
Federico Martelli1 месяц назад

@PantheonInc catching swapped camera streams and mislabeled instructions before training saves more than another round of fine tuning, factory data has the same mess

Фото профиля Alexander
Alexander1 месяц назад

@PantheonInc Super cool ! Have you guys looked at how much variation there is in SR of VLA style models when there is variance in the task description?

Фото профиля ^_^
^_^1 месяц назад

@PantheonInc That's amazing!

Фото профиля Jack
Jack1 месяц назад

Generated annotations need a sampling audit or you inherit the model's blind spots at full scale. The failure mode is not random noise, it is consistent error, which looks exactly like clean data until something downstream learns the same mistake. Worth holding back a human labelled slice purely to measure drift against.

Фото профиля Anna /✦
Anna /✦1 месяц назад

@PantheonInc sharing with my friend in robotics!

Фото профиля Kifayath Ulla
Kifayath Ulla1 месяц назад

@PantheonInc WOAHH

Похожие видео

🔥 JUST IN: Open-source robotics dataset from 100% real-world scenarios! 🤯 Chinese robotics company AGIBOT just released AGIBOT WORLD 2026, an open-source dataset systematically covering key embodied AI research directions. Built entirely from real-world environments: commercial spaces, and homes. Collected using AGIBOT G2 robots in free-form collection mode, providing structured, accurately annotated, high-quality data. Digital twin technology creates 1:1 scale replicas in simulation matching the real environments. Both real-world and simulation data are open-sourced. The AGIBOT G2 platform collects multiple data types simultaneously: RGB(D) cameras, tactile sensors, force sensors, LiDAR, IMU, and full-body joint states. Whole-body control coordinates arms, waist, and hands for complex tasks. First-person teleoperation lets operators control the robot from its perspective. The tasks covered are fine-grained manipulation, ultra-long-horizon tasks, spatial navigation, dual-arm coordination, and multi-agent/human-robot collaboration. The dataset includes error-recovery trajectories with annotations. Most datasets only show successful demonstrations. AGIBOT includes failures and how the robot recovers, teaching models how to handle mistakes. After collection, data is tested through policy training and real-robot deployment to ensure quality. Then processed through industrial quality control with multiple screening and cleaning rounds. Making it open-source accelerates embodied AI research by giving researchers access to high-quality real-world robot data at scale. 🇨🇳 Learn more here: ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

Lukas Ziegler

40,583 просмотров • 7 месяцев назад