Lumo-2's Three-Stage Training Method
rohanpaul_ai · x · 2026-07-18
This section clarifies the training pipeline of Lumo-2:
- Stage 1: Align visual changes with robot actions.
- Stage 2: Give action encodings semantics through vision and language.
- Stage 3: Jointly train VLM, video, and robot data, so the model predicts the next physical state before outputting actions.
The overall idea is to compress "how the world will change" into a small latent state, which then guides control.
Related event: Astribot launches Lumo-2 with real-robot demos(10 posts)→
More from Embodied
- RTX Spark laptops impress analyst: 30B local model runs quiet with great battery life — Scobleizer · 2026-10-27
- Re-uploading a 239GB robot dataset moved only 2.6MB — inside a petabyte-scale data pipeline — mervenoyann · 2026-10-27
- ROSCon UK photo tour: BBC's whale robot, da Vinci surgical robot and the OG PR2 — kamathsblog · 2026-10-27
- Teardown of a humanoid robot Z-axis linear actuator, with Zmotor Q&A — MarwaEldiwiny · 2026-10-27
- GPT-6 Astra Beats Gemini 3.8 Flash in Paired Robot Test, Yet Neither Completes the Task — qinzytech · 2026-10-27
- Teardown of Zmotor LA500 humanoid linear actuator: 324 g, 500 N peak thrust — MarwaEldiwiny · 2026-10-27