8000 How to modify the task language in LIBERO to test the ability of OpenVLA to adapt to different language inputs? · Issue #79 · moojink/openvla-oft · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content
How to modify the task language in LIBERO to test the ability of OpenVLA to adapt to different language inputs? #79
Open
@ziyangc2

Description

@ziyangc2

Hi, I am testing OpenVLA-OFT using experiments/robot/libero/run_libero_eval.py. BUT when I modify task_description by code below:

originally task description is: put both the alphabet soup and the tomato sauce in the basket

env, task_description = get_libero_env(task, cfg.model_family, resolution=cfg.env_img_res) task_description = "Put milk in the busket"

I also edited bddl file of this task, but there is no difference in rendered video.

The video I rendered is the same as the one before modifying the input task description. How can I test the ability of OpenVLA-OFT to adapt to different language inputs by modifying the task descriptions in LIBERO?

Thanks for any ideas and suggestions.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions

      0