v0.12.0

What's Changed

• fixed equiv mp training as shallow ensemble by @M-R-Schaefer in #441
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #440
• Transfer Learning Freeze layers by @M-R-Schaefer in #443
• removed buggy n jit steps option by @M-R-Schaefer in #446
• orbax checkpoint saving / loading by @PythonFZ in #445
• test selection on atoms without calc by @M-R-Schaefer in #447
• add ApaxApplyTransformation by @PythonFZ in #433
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci in #452
• bump version to 0.12.0 by @M-R-Schaefer in #453

Full Changelog: v0.11.0...v0.12.0