AME2 neural mapping adapted for Livox lidar trained within 1 day
ChongZzZhang · x · 2026-08-17
Demonstrated the adaptation of Spare-time-vibe-adapted ame2 neural mapping for a g1 robot with Livix lidar. A 2k-iteration student model was trained from a 3k-iteration teacher. The entire training process completed in one day, including 1.5 hours for neural mapping. The model shows accurate predictions but is slightly underconfident.
Related event: Robot Policies Trained in a Day via Massive Parallel Sims and Lidar Mapping(2 posts)→
More from Embodied
- Tesla FSD v14.3.6 Revealed: 10B MoE Model and RL Enhancements — qinzytech · 2026-08-17
- HopTo's Hop-1 Robot Uses Hopping to Cut Energy Use by 75% — MarwaEldiwiny · 2026-08-17
- New robot training method: taga-style attention enables massive parallelization — ChongZzZhang · 2026-08-17
- FCC adds foreign ground robots to Covered List; clarifies it's not a ban and targets production location — the-uncanny-squad · 2026-08-17
- Hosted Micro Robot Learning Hackathon at Office Over the Weekend — DominiqueCAPaul · 2026-08-17
- Virtuals Weekly: Eastworlds Becomes Unitree's Official Data Partner, Largest G1 Data Source Outside China — DonJohnsonSays · 2026-08-17