0.0.5

Full Changelog: 0.0.4...0.0.5

0.0.4

What's Changed

• adding dmc quadruped env by @quentinll in #130
• [BugFix] fix tworoom by @lucas-maes in #131
• [BugFix] evaluate_from_dataset variation by @lucas-maes in #133
• [Env] two-room by @lucas-maes in #134
• New tworoom by @lucas-maes in #135
• Add rendering option for target visibility in DMControl environments by @quentinll in #136
• [BugFix] Clip idx in dataset by @lucas-maes in #137
• [Feat] expert policy for DMC by @lucas-maes in #138
• [Docs] update for DMC expert by @lucas-maes in #139
• [Feature] Add Gif by @lucas-maes in #140
• script: add DMC data collection scripts by @lucas-maes in #142
• Adding tasks for DMC envs by @quentinll in #143
• Fixing GCBC and IQL baselines by @quentinll in #132
• feat: add col dim in dataset by @lucas-maes in #145
• fix dataset for np.object_ by @lucas-maes in #146

Full Changelog: 0.0.1...0.0.4

0.0.2

Big refactor, better dataset, docs, and tests.

0.0.1

What's Changed

• [Feature] Add stacked wrapper by @lucas-maes in #53
• add goal info in first_steps by @lucas-maes in #55
• Update GitHub Actions workflow permissions by @lucas-maes in #62
• Refactor StackedWrapper to support multiple keys by @quentinll in #61
• [BugFix][Feature] Stacked Wrapper by @lucas-maes in #65
• Improving CEM implementation by @quentinll in #63
• Reproducible CEM by @quentinll in #67
• [BugFix] Correct dino-wm training and evaluation by @lucas-maes in #72
• Bump version to 0.0.1b1 by @lucas-maes in #73

Full Changelog: v0.0.1b0...v0.0.1b1

v0.0.1b0

What's Changed

• fix slow data loading by @lucas-maes in #21
• dino-wm script by @lucas-maes in #23
• dino-wm by @lucas-maes in #24
• Faster data by @lucas-maes in #25
• Major Patch (working dino-wm, refactor dataset recording) by @lucas-maes in #31
• OGBench Cube Env Addition by @DanHrmti in #28
• Add integration test for new world + doc by @lucas-maes in #35
• Add multi-view support for pixels by @lucas-maes in #34
• add object_dump_callback by @lucas-maes in #37
• The finished running script for training dino-wm by @lucas-maes in #38
• OGBench Scene Env by @DanHrmti in #33
• Delete .hydra directory by @lucas-maes in #42
• Clean up dependencies section in pyproject.toml by @lucas-maes in #45
• Fix dinowm script by @quentinll in #41
• Remove stable-pretraining from dependencies by @lucas-maes in #46
• Pusht variations by @quentinll in #44
• Replay evaluator (offline) by @lucas-maes in #43
• Preempted Inference tmp file logic by @lucas-maes in #49
• add seed for dinowm data split by @lucas-maes in #50
• better print variation space by @lucas-maes in #51
• fix versioning by @lucas-maes in #52

New Contributors

• @DanHrmti made their first contribution in #28
• @quentinll made their first contribution in #41

Full Changelog: https://github.com/rbalestr-lab/stable-worldmodel/commits/v0.0.1b0