smolvla-vlabench

Automate Franka Panda pick, place, and grasp tasks with Vision-Language-Action policies.

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

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill provides a Vision-Language-Action policy that automates robotic manipulation tasks, enabling precise actions like picking, placing, and grasping objects on a tabletop.

Core Features & Use Cases

  • Vision-Language-Action (VLA) Policies: Utilizes VLA policies for understanding natural language descriptions and performing robotic tasks.
  • Embodied on Franka Panda: Specifically designed to work with the Franka Panda robotic arm.
  • VLABench Integration: Integrated with the VLABench robotic manipulation benchmark for testing and validation.
  • Use Case: Imagine you have a robot that needs to pick up a fruit from a table, place it into a bowl, and then grasp an object. Use this Skill to program the robot's actions using natural language descriptions.

Quick Start

Run the following command to execute the policy: openral benchmark scene --config scenes/benchmark/vlabench_select_fruit.yaml --rskill rskills/smolvla-vlabench

Frequently Asked Questions about smolvla-vlabench

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

FAQPage Schema
How do I automate robotic manipulation tasks with vision-language-action policies?

You can automate robotic manipulation tasks using vision-language-action policies to execute precise picking, placing, and grasping actions. The system interprets natural language descriptions to control the Franka Panda robotic arm on a tabletop.

What is a vision-language-action policy for robotic manipulation?

A vision-language-action policy for robotic manipulation is a model that understands natural language descriptions and translates them into precise physical actions. It enables robots to perform tasks like picking and placing objects based on visual input and text commands.

Does the Franka Panda robot support natural language commands for grasping objects?

Yes, the Franka Panda robot supports natural language commands for grasping objects through integrated vision-language-action policies. You can program the robot to pick up items like a fruit and place them into a bowl using text descriptions.

Do I need PyTorch and OpenRAL to run vision-language-action benchmarks?

Yes, you need PyTorch and OpenRAL to run vision-language-action benchmarks. PyTorch provides the required runtime environment, while the OpenRAL framework is necessary for executing the robotic manipulation policies and scenes.

What are the limitations of using vision-language-action policies for tabletop manipulation?

Limitations of using vision-language-action policies for tabletop manipulation include being specifically designed for the Franka Panda robotic arm and requiring integration with the VLABench benchmark for robustness testing. It also depends on PyTorch and OpenRAL environments.