Skip to content

[RLlib] TFLearner does not correctly implement restore_from_path #49860

Open
@Mark2000

Description

What happened + What you expected to happen

When restoring a checkpoint with TFLearner, it fails on restore_from_path because it does not pass the component argument through to the superclass.

Versions / Dependencies

Present on current Ray release (2.40.0)

Reproduction script

Call Algorithm.from_checkpoint on a DreamerV3 checkpoint.

Issue Severity

High: It blocks me from completing my task.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Assignees

No one assigned

    Labels

    P1Issue that should be fixed within a few weeksbugSomething that is supposed to be working; but isn'trllibRLlib related issuesrllib-checkpointing-or-recoveryAn issue related to checkpointing/recovering RLlib Trainers.rllib-newstack

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions