Meta Platforms Inc.Meta's open-source emg2pose dataset is used by Wetour Robotics, showcasing its value for third-party innovation.

Wetour Robotics has released a live demonstration of its Conductor neural wristband, which decodes muscle signals into a real-time 3D hand digital twin and converts gestures into text commands without cameras or gloves. The sEMG wristband uses an 8-channel sensor at 250 hertz, and the company trained its model on Meta’s openly released emg2pose dataset, downsampled to match Conductor’s consumer-grade hardware, then fine-tuned it with transfer learning on its own data. The streaming state-space model runs fully on-device, and the demo evaluated it on gestures it had not previously seen. Conductor is part of Wetour’s open, cross-device Orchestra platform, positioned as a human-intent data layer for robotics and future physical-world models, with an enterprise Early Access Program now open to qualified partners.
Meta Platforms Inc.Meta's open-source emg2pose dataset is used by Wetour Robotics, showcasing its value for third-party innovation.