smolvla-libero

Execute SmolVLA vision-language-action policies for robotic manipulation tasks.

Updated Jul 5, 2026
One-click install
npx skills add https://github.com/bensonlee5/openral --skill smolvla-libero
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: smolvla-libero
Source: https://github.com/bensonlee5/openral/tree/main/rskills/smolvla-libero
Command: npx skills add https://github.com/bensonlee5/openral --skill smolvla-libero

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires openral, pytorch, huggingface_hub, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill provides an efficient Vision-Language-Action policy for robotic manipulation tasks, enabling robots to perform tasks like picking up objects, placing them, and manipulating them in a controlled manner.

Core Features & Use Cases

  • Vision-Language-Action: Combines vision, language, and action to enable robots to understand and execute tasks described in natural language.
  • SmolVLA Model: Utilizes the SmolVLA model, which is finetuned on the LIBERO task suite for efficient performance.
  • Embodiment Support: Supports various robot embodiments, including Franka Panda, SO-100 follower arm, and any 6-7 DOF manipulator.
  • Use Case: Imagine you have a robot equipped with a camera and a gripper. Use this Skill to instruct the robot to pick up a red cube from a table and place it in a drawer using natural language commands.

Quick Start

Use the rskill-smolvla-libero skill to perform a task on the Franka Panda robot.

Frequently Asked Questions about smolvla-libero

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I execute robotic manipulation tasks using natural language commands?

Robotic manipulation tasks using natural language commands are executed via a Vision-Language-Action policy. This Skill uses the SmolVLA model to translate text instructions into physical actions, enabling robots to pick and place objects.

Does this Vision-Language-Action policy support different robot embodiments?

Yes, the Vision-Language-Action policy supports various robot embodiments. It is compatible with the Franka Panda, SO-100 follower arm, and any 6-7 DOF manipulator through the OpenRAL loader and robot HAL.

What hardware do I need to run the SmolVLA model for robotic manipulation?

To run the SmolVLA model for robotic manipulation, you need a robot embodiment equipped with cameras and grippers. These components are required for the Vision-Language-Action policy to perceive and interact with the environment.

How do I get started with a SmolVLA model on the Franka Panda robot?

To get started with the SmolVLA model on the Franka Panda robot, use the provided scripts to execute the Skill. The model is finetuned on the LIBERO task suite and runs via the OpenRAL loader and robot HAL.

What is the LIBERO task suite used for in Vision-Language-Action models?

The LIBERO task suite is used to finetune Vision-Language-Action models like SmolVLA. It provides a set of robotic manipulation tasks to train the policy for efficient performance in picking, placing, and object manipulation.

Are there limitations when using a Vision-Language-Action policy for object manipulation?

A limitation of using a Vision-Language-Action policy for object manipulation is the strict hardware requirement. You must have compatible cameras and grippers attached to a 6-7 DOF manipulator for the policy to function.