Project-IRA/TPSoSe2026_SmolVLA_LeRobot_SO101_Finetuning_V2_Lego

Robotics Model

This is a fine-tuned SmolVLA vision-language-action model for the SO-101 robot, trained on a Lego manipulation dataset using LeRobot and flow-matching, with a quality score of 0.8571.

Catalog metadata

Repository owner: Project-IRA
Source: Hugging Face Models
Task: robotics
Cameras: 0
License: cc-by-sa-4.0

Browse the catalog to compare robots, tasks, sensors, and licenses. Sign in to open protected download and model links.