Frozen Qwen2.5-VL Robotics Policy

This submission contains a local frozen Qwen2.5-VL-3B-Instruct backbone and a trained 7D action head. The adapter consumes only the documented image, instruction, task, difficulty, proprio, step, and horizon observation fields.

Total parameter count: 3,755,818,503.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support