you balanced your robot dataset by embodiment. the embeddings say you balanced it by recording session i pulled 10 episodes from each of the 50 robot types in lerobot's community dataset and embedded every clip with qwen3-vl 1. 495 of 497 clips have their nearest neighbor in the same session 2. 292 score below 0.1 uniqueness 3. the two "most unique" clips are single-frame recordings, 0.02 seconds long filter the session out and the cross-embodiment signal shows up: a cloth fold on one rig lands next to so100 arms folding cloth i indexed all 497 episodes in fiftyone so you can search by text, cut the near-duplicates, and export a curated lerobot v3 subset one prompt to an agent with fiftyone skills built the whole thing start here, read the full blog: https://huggingface.co/blog/harpreetsahota/fiftyone-now-reads-lerobot-50-embodiments-497-epis read the dataset card: https://huggingface.co/datasets/Voxel51/community_v3_10per_embodiment @LeRobotHF @Alibaba_Qwen
Open Artifacts
AllTopicsOpen Artifacts with links to supporting source material.
Latest Signal
No. 1–20 records · Total 20 records
LeRobot@LeRobotHFSignalEditorial score6565Quoted postharpreet@DataScienceHarpChinese robotics — Dataset and collection discoverySignalEditorial score8585 Teaching Robots to Work: A Dataset and Collection Discovery in Chinese Robotics
At the 2026 China International Digital Economy Expo, a demonstration showcased how robots are being trained using real-world data collected through VR and exoskeletons. The article details the construction of over 70 embodied intelligence training facilities, the challenges of data collection and annotation, and the role of platforms like Qingyan Tech in data governance and trading. The focus is on the infrastructure and ecosystem supporting the AI robotics industry.
Machine Heart — Robotics on WeChatSignalEditorial score8585 CoRL 2026|Making Robots More Agile: Starting with Learning to Walk
This paper presents LeaP, a learnable source prior for generative robot policies, which replaces fixed standard Gaussian distributions with state-dependent random initialization. LeaP jointly learns the mean and variance of an initial Gaussian distribution, improving action generation in robotic tasks. The method achieves an average success rate of 81.6% in simulation and 80.0% on real robotic arms, outperforming existing baselines like A2A and VITA. The paper is accepted at CoRL 2026, and the code is open-sourced.
International Journal of Robotics Research — Journal metadataSignalEditorial score8585 GrandTour: A legged robotics dataset in the wild for multi-modal perception and state estimation
The GrandTour dataset addresses the lack of large-scale, real-world legged-robot data by providing time-synchronized sensor data from multiple modalities, including LiDAR, RGB cameras, depth sensors, and GNSS. It supports research in SLAM, state estimation, and multi-modal learning, with data available in ROS and HuggingFace formats.
Animesh Garg@animesh_gargSignalEditorial score8585Quoted postBlack Forest Labs@bfl_aiIntroducing FLUX 3 Action. An open weights 7B World Action Model that achieves first place on the RoboLab benchmark. It outperforms the previous best open model by 6.1 percentage points while using 56% fewer parameters and running up to 3.95x faster. FLUX 3 Action removes the usual trade-off between world action model performance and VLA speed: it still predicts video and actions together, but plans more than twice as far ahead and runs faster per second of robot motion than the strongest open VLA. Teams can fine-tune FLUX 3 Action on their own demonstrations to create policies for a particular robot and task. Together with @nvidia, we also integrated FLUX 3 Action natively into @huggingface's LeRobot, with fine-tuning recipes included and edge deployment on NVIDIA Jetson. Beyond robotics, we’re also seeing promising results training task-specific policies for acting in simulated environments like gaming, controlling a vehicle, computer use, and wherever else a model needs to understand a visual environment and then choose what to do next. FLUX 3 Action builds on the same image, video, and audio pretraining as FLUX 3, but uses a smaller architecture designed for practical deployment. In midtraining, we trained the model to predict actions and future frames together. We’re releasing the weights, code, fine-tuning recipe, benchmarks, and reproducible examples so researchers and developers can build on the model with their own robots, environments, and tasks (see below).
Robotics Update — UK and European industrySignalEditorial score8585 Universal Robots unveils next generation platform
Universal Robots has launched Gen 7, its seventh-generation robotics platform, featuring a reengineered control box, teach pendant, and tool flange. The platform includes new g-Series robot arms, the CB7 Core controller, PolyScope X operating system, and TP7 teach pendant, all aimed at improving AI integration, safety, and ease of use in industrial automation.
Universal Robots — Official announcementsSignalEditorial score8585 Universal Robots unveils Gen 7, a new platform for industrial automation and physical AI deployment
Universal Robots launched Gen 7, its seventh-generation robotics platform, featuring AI-ready hardware, a reengineered control box, and PolyScope X software. The platform aims to simplify AI deployment in industrial automation with enhanced safety, cybersecurity, and open architecture.
Unitree@UnitreeRoboticsSignalEditorial score6363
NVIDIA — RoboticsSignalEditorial score8888 Into the Omniverse: How Open World Models Push the Frontier of Physical AI
NVIDIA introduces Cosmos 3, an open-world model family for physical AI, combining vision reasoning, world generation, and action prediction. Available under the Linux Foundation’s OpenMDW 1.1 license, it enables post-training and adaptation for robotics, autonomous vehicles, and vision AI systems. The model ranks highly in benchmark evaluations and is part of NVIDIA’s broader physical AI stack.
LeRobot — Model repository updatesSignalEditorial score6565 Hy-Embodied-0.5-VLA — LeRobot Conversion
LeRobot has released a converted version of the Hy-Embodied-0.5-VLA-RoboTwin model, achieving full forward parity with the original model. The model supports a 16D dual-arm embodiment and has been validated across multiple metrics, including task completion and action accuracy. The release highlights the model's performance on a RoboTwin quick protocol and notes limitations in expressing mobile-manipulator actions.
Robotiq — BlogSignalEditorial score6565 Tactile-Based Robot Centering as a Capability for Dexterous Manipulation
This article explains how tactile sensing allows robots to detect and correct off-center grasps by analyzing force distribution across fingertips. It describes a demonstration using a 2F-85 gripper with TSF-85 tactile sensors, which repositions and regrasps objects to achieve a balanced grip. The method improves grasp stability and reduces the need for reorientation or regrasping.
NVIDIA — RoboticsSignalEditorial score8585 NVIDIA Open Sources First GPU-Accelerated Medical Physics Simulation Framework
NVIDIA has released an open-source, GPU-accelerated Medical Physics Simulation framework as part of its Isaac for Healthcare platform. This tool enables developers to model anatomy-device interactions, generate complex scenarios, and train robot policies in simulation, significantly reducing development time and improving regulatory readiness.
NVIDIA — RoboticsSignalEditorial score8585 NVIDIA and Hugging Face Bring New Models and Frameworks to LeRobot for the Open Robotics Community
NVIDIA and Hugging Face are releasing the NVIDIA Isaac GR00T 1.7 model and Isaac Teleop framework into LeRobot, an open-source robotics library, to provide developers with shared tools for training, evaluating, and deploying robot foundation models. NVIDIA Cosmos 3, a frontier world model for physical AI, is also set to be integrated soon.
Unitree@UnitreeRoboticsSignalEditorial score6565Quoted postBitRobot 🦾@BitRobotNetwork1/ Introducing HIW-500 (Humanoids-in-the-Wild 500): the largest open-source humanoid teleop dataset collected in real homes Built w/ @UnitreeRobotics @huggingface across 12 homes in Southeast Asia, it covers: > 500+ hrs > 23K+ episodes > 10+ TB > 10+ household tasks
PickNik — Robot softwareSignalEditorial score6565 MoveIt Pro Newsletter June 2026
PickNik Robotics announces the release of MoveIt Pro, a commercial robotics software platform that extends the open-source MoveIt project. The update includes support for AI models like Gemini, SAM3, and Codex, as well, enhancements to visual reasoning, trajectory stitching, and ROS compatibility. The platform is designed for unstructured environments and supports a wide range of robotic arms, t
LeRobot — Model repository updatesSignalEditorial score8888 LeRobot: π₀.₅ Model Release for Vision-Language-Action Tasks
LeRobot releases π₀.₅, a Vision-Language-Action (VLA) model with open-world generalization capabilities, trained on robot demonstrations and multimodal data. The model supports flow-matching for action prediction and is intended for fine-tuning on specific use cases. It is available for inference and training with detailed documentation and example code.
PickNik — Robot softwareSignalEditorial score6565 MoveIt Pro Newsletter September 2025
PickNik announces the release of MoveIt Pro, a commercial robotics software platform built on the open-source MoveIt project. The update highlights new features such as support for FANUC and Kawasaki robots, integration with Meta Quest for AI data collection, Diffusion Policies for end-to-end AI training, and enhanced simulation capabilities with IMU and 3D Lidar support. The platform is designed for unstructured environments, providing tools for motion planning, grasping, perception, and behavior sequencing.
Hugging Face — RoboticsSignalEditorial score8383 Post-Training Isaac GR00T N1.5 for LeRobot SO-101 Arm
NVIDIA has released the GR00T N1.5 model, a cross-embodiment foundation model for generalized humanoid robot reasoning and skills. The model can be fine-tuned using teleoperation data from a SO-101 arm, with a detailed tutorial provided for developers. The release includes instructions for dataset preparation, fine-tuning, evaluation, and deployment.
Hugging Face — RoboticsSignalEditorial score8888 SmolVLA: Efficient Vision-Language-Action Model trained on Lerobot Community Data
Hugging Face introduces SmolVLA, a 450M parameter open-source Vision-Language-Action model for robotics, trained on community-shared datasets. It outperforms larger models in simulation and real-world tasks, supports asynchronous inference for faster response, and is designed for deployment on consumer hardware.
Open Robotics — BlogSignalEditorial score6565 ROS 2 Kilted Kaiju Released
Open Robotics announces the release of ROS 2 Kilted Kaiju, a new version of the Robot Operating System with support for Zenoh middleware, Python improvements, and enhanced ROSBag features. The release is supported on Ubuntu 24.04 and Windows, with plans for future improvements in the upcoming Lyrical Luth release.