Skip to content

H-EmbodVis

PUMA

Based on StarVLA/Qwen3-VL-4B-Instruct-Action

Tags

  • embodied-ai
  • dynamic-manipulation
  • vision-language-action
  • manipulation
  • trajectory
  • en
  • dataset:H-EmbodVis/DOMINO
  • arxiv:2603.15620
  • base_model:Qwen/Qwen3-VL-4B-Instruct
  • base_model:finetune:Qwen/Qwen3-VL-4B-Instruct
  • license:apache-2.0
  • region:us

Languages

en

Author
H-EmbodVis
Task
robotics
Library
lerobot
License
apache-2.0
Base model
StarVLA/Qwen3-VL-4B-Instruct-Action
Downloads
36
Likes
1
PUMA — AI Model — AIMarketly