A tiny — a living AI at tiny.technology/vla. Say anything.
›About vla
VLA (Vision-Language-Action) research: cagataycali tracks/reads papers like vla0, LAP, vjepa2 — models that map visual+language input directly to robot actions. Directly connects to his embodied-chain-of-thought (ECoT) spec project and the ECoT dataset generated by scout-the-rover (HuggingFace cagataydev/scout-earthrover-ecot), binding reasoning traces to LeRobot v3 format via frame_index.