Skip to content

Pi0.5 reproduction #973

@SMLee0407

Description

@SMLee0407

Hello,

I’m trying to fine-tune Pi0.5 on LIBERO or RoboCasa, but it seems that this repo only provides a pretrained Pi0.5 checkpoint and does not include subtask prediction.

Is it typical to reproduce Pi0.5 using the same fine-tuning process as Pi0, with only the continuous action prediction loss for the action expert and VLM?

In other words, is the checkpoint path the only difference between reproducing Pi0 and Pi0.5?

Thanks.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions