Bump sb3-contrib from 2.7.1 to 2.8.0

[dependabot]

Bumps sb3-contrib from 2.7.1 to 2.8.0.

Release notes

Sourced from sb3-contrib's releases.

v2.8.0: Fix for MaskablePPO and markdown doc

Breaking Changes:

• Removed support for Python 3.9, please upgrade to Python >= 3.10
• Upgraded to Stable-Baselines3 >= 2.8.0
• Set strict=True for every call to zip(...)

New Features:

• Added official support for Python 3.13

Bug Fixes:

• Fixed MaskablePPO and RecurrentPPO inaccurate n_updates counting when target_kl early exits the training loop
• Fixed RecurrentPPO and MaskablePPO forward and predict not reshaping the action before clipping it (@​immortal-boy)
• Do not call forward() method directly in RecurrentPPO (@​immortal-boy)
• Fixed MaskableCategorical.apply_masking() crashing with ValueError: Simplex when cached probs deviate from sum=1 in float32 with large action spaces (torch 2.9+) (@​kirann-05)

Documentation:

• Switched to markdown documentation (using MyST parser)

New Contributors

• @​immortal-boy made their first contribution in Stable-Baselines-Team/stable-baselines3-contrib#320
• @​kirann-05 made their first contribution in Stable-Baselines-Team/stable-baselines3-contrib#326

Full Changelog: Stable-Baselines-Team/stable-baselines3-contrib@v2.7.1...v2.8.0

Commits

• 50d8a8f Release v2.8.0 (#328)
• 300a5ed fix: clear cached probs in MaskableCategorical.apply_masking to avoid Simplex...
• 6fa2410 Update contribution guidelines regarding LLM/code assistant usage (#327)
• 75fff3a Fix n_updates logging in MaskablePPO and RecurrentPPO when using early stop...
• ee23ad3 doc: switch to Markdown (MyST parser) (#324)
• e03aa1e Add support for multi-dimensional action spaces to PPO variants (#318)
• 06c6e62 Remove explict forward calls in RecurrentPPO (#320)
• d67de80 Add Python 3.13 support, drop Python 3.9 (#319)
• See full diff in compare view