update rsl_rl to 3.1.2 to support state-dependent std dev

[ashwinvkNV]

Description

Updating rsl_rl to 3.1.2 release to support use these two changes in IsaacLab sim-to-real deployed policies.

Main commits of interest:
leggedrobotics/rsl_rl@530f71a
leggedrobotics/rsl_rl@a4d108a

Type of change

• New feature (non-breaking change which adds functionality)

• I have read and understood the contribution guidelines

• I have run the pre-commit checks with ./isaaclab.sh --format

• I have made corresponding changes to the documentation

• My changes generate no new warnings

• I have updated the changelog and the corresponding version in the extension's config/extension.toml file

• I have added my name to the CONTRIBUTORS.md or my name already exists there

[greptile-apps]

Greptile Overview

Greptile Summary

This PR updates the RSL-RL library dependency from version 3.0.1 to 3.1.2 across two files to enable state-dependent standard deviation support for RL policies. The changes include updating the pinned version in source/isaaclab_rl/setup.py (line 49) and updating the minimum version constant in scripts/reinforcement_learning/rsl_rl/train.py (line 59). This upgrade is necessary for sim-to-real deployment of manipulation policies in IsaacLab, as referenced in the PR description. The existing version validation infrastructure in the training script will automatically enforce this requirement and guide users to upgrade if needed. The changes are minimal and focused entirely on version numbers, leveraging the codebase's existing RSL-RL integration infrastructure including policy exporters, wrappers, and configuration classes.

PR Description Notes:

• Minor typo: "support use these two changes" should likely be "support these two changes" or "use these two changes"

Important Files Changed

Filename	Score	Overview
scripts/reinforcement_learning/rsl_rl/train.py	5/5	Updated RSL_RL_VERSION constant from "3.0.1" to "3.1.2" to enforce minimum version requirement
source/isaaclab_rl/setup.py	4/5	Updated rsl-rl-lib dependency from version 3.0.1 to 3.1.2 in the package requirements

Confidence score: 4/5

• This PR is safe to merge with low risk as it contains only version string updates
• Score reflects that while the changes are minimal and well-isolated, the PR checklist indicates the changelog and extension.toml version were not updated (which may be required by contribution guidelines), and there is no explicit verification that rsl-rl 3.1.2 maintains backward compatibility with existing IsaacLab RSL-RL integration code
• Pay close attention to source/isaaclab_rl/setup.py to ensure the version bump does not introduce breaking changes in RSL-RL wrappers, exporters, or configuration classes

_{2 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

Greptile Overview

Greptile Summary

Updates the rsl-rl-lib dependency from version 3.0.1 to 3.1.2 to support state-dependent standard deviation and better handling of std_cond_in parameters for sim-to-real deployment policies.

Changes:

• Updates dependency version in source/isaaclab_rl/setup.py from rsl-rl-lib==3.0.1 to rsl-rl-lib==3.1.2
• Updates minimum version check in scripts/reinforcement_learning/rsl_rl/train.py from 3.0.1 to 3.1.2

Key Benefits:

• Adds support for state-dependent standard deviation in actor networks
• Enables better handling of conditional input parameters for policy deployment
• Aligns with upstream rsl_rl improvements needed for IsaacLab sim-to-real policies

Confidence Score: 5/5

• This PR is safe to merge with minimal risk - it's a straightforward dependency version bump
• The changes are minimal and focused: only updating version numbers in two locations (setup.py dependency and train.py version check). The update enables new features (state-dependent std deviation) that are backward-compatible with existing code. The exporter and wrapper code already handle the actor network structure generically, so no breaking changes are expected.
• No files require special attention

Important Files Changed

File Analysis

Filename	Score	Overview
source/isaaclab_rl/setup.py	5/5	Updates rsl-rl-lib dependency from 3.0.1 to 3.1.2
scripts/reinforcement_learning/rsl_rl/train.py	5/5	Updates minimum required rsl-rl version constant from 3.0.1 to 3.1.2

Sequence Diagram

sequenceDiagram
    participant User
    participant TrainScript as train.py
    participant Setup as setup.py
    participant RslRl as rsl-rl-lib
    participant ActorNet as ActorCritic
    
    User->>TrainScript: Execute training
    TrainScript->>TrainScript: Check version >= 3.1.2
    alt Version check fails
        TrainScript-->>User: Error: Upgrade required
    else Version check passes
        TrainScript->>Setup: Load dependencies
        Setup->>RslRl: Import rsl-rl-lib 3.1.2
        RslRl-->>TrainScript: Library loaded
        TrainScript->>RslRl: Create OnPolicyRunner
        RslRl->>ActorNet: Initialize with state-dependent std
        ActorNet-->>RslRl: Policy ready
        RslRl-->>TrainScript: Runner initialized
        TrainScript->>RslRl: Start training
        RslRl->>ActorNet: Train with std_cond_in support
        ActorNet-->>RslRl: Updated weights
        RslRl-->>TrainScript: Training complete
        TrainScript-->>User: Export for deployment
    end

Loading

_{2 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}