Manipulation datasets for SAGE: Subgoal-Conditioned Action Generation for Latent World Model Planning. Sizes are archive sizes in decimal GB. Training and validation archives are under training/; evaluation archives are under evaluation/.
Dataset Card
Manipulation datasets for SAGE: Subgoal-Conditioned Action Generation for Latent World Model Planning. Sizes are archive sizes in decimal GB. Training and validation archives are under training/; evaluation archives are under evaluation/. Evaluation archives contain the inputs required by the released query manifests, not a complete training dataset. Simulator assets and model weights are installed separately. Download only the evaluation data you need, for example LIBERO Scene2: To download Scene2 training and validation instead, use allowpatterns=["training/liberoscene2"]. Replace liberoscene2 with liberocaddy or robotwina2b for the other environments. Omitting allowpatterns downloads all…
Excerpt from the card by Letian Cheng.
Details
- Repository
- CLTRAY/SAGE-data
- Publisher
- Letian Cheng
- Task category
- Robotics
- Tags
- robotics, world-model, planning
- Size category
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- 2431d40a69530efc0796ab9f2e532d0dfa91595e
- Last updated
- 2026-10-14
Files
59 files, 285.2 GB in total.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| evaluation/libero_caddy_evaluation.tar | Data | 8.6 GB | 0b5c7b6d0305 |
| evaluation/libero_scene2_evaluation.tar | Data | 19.3 GB | 334eefe83e88 |
| evaluation/robotwin_a2b_evaluation.tar | Data | 1.4 GB | 66beba192246 |
| runtime/robotwin/robotwin_a2b_runtime.tar.gz | Data | 48.1 KB | 71f68ca0fc6d |
| training/libero_caddy_train.tar.parts.json | Data | 1.7 KB | 5a6c77ae27ce |
| training/libero_caddy_validation.tar | Data | 8.6 GB | 3bafb6608075 |
| training/libero_scene2_train.tar.parts.json | Data | 2.1 KB | b687a306ccf9 |
| training/libero_scene2_validation.tar | Data | 9.9 GB | b6fca0de3e04 |
| training/robotwin_a2b_train.tar.parts.json | Data | 1.9 KB | 39c2081a53bb |
| training/robotwin_a2b_validation.tar | Data | 9.0 GB | fa66856b89eb |
| README.md | Documentation | 3.8 KB | — |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part000 | Other | 536.9 MB | eac47001d68e |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part001 | Other | 536.9 MB | 2ee598d3d9dc |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part002 | Other | 536.9 MB | 690890739297 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part003 | Other | 536.9 MB | 98aaaa4554b9 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part004 | Other | 536.9 MB | 98f3b0f303ff |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part005 | Other | 536.9 MB | 0f39895a9ed0 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part006 | Other | 536.9 MB | e502c0aae86f |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part007 | Other | 536.9 MB | c1a94fc51e02 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part008 | Other | 536.9 MB | f964129a1c26 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part009 | Other | 536.9 MB | 1b00e29d51b9 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part010 | Other | 536.9 MB | df3c4b79f087 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part011 | Other | 536.9 MB | 8d3fb58f8d22 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part012 | Other | 536.9 MB | 99e8e9dd95e7 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part013 | Other | 536.9 MB | 0c3bcb7d10db |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part014 | Other | 536.9 MB | e9faec1d78aa |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part015 | Other | 536.9 MB | 67ddad788fed |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part016 | Other | 536.9 MB | 3cd2be6185f8 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part017 | Other | 536.9 MB | c7f0166fc51e |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part018 | Other | 536.9 MB | 318d33a495f4 |
| runtime/robotwin/robotwin_a2b_assets.tar.gz.part021 | Other | 325.1 MB | 0bb623c50bfd |
| training/libero_caddy_train.tar.part001 | Other | 8.6 GB | 1db40552a3c4 |
| training/libero_caddy_train.tar.part002 | Other | 8.6 GB | 0d0d5ba1a943 |
| training/libero_caddy_train.tar.part003 | Other | 8.6 GB | e341d68ed0f9 |
| training/libero_caddy_train.tar.part004 | Other | 8.6 GB | 367139eb8a5a |
| training/libero_caddy_train.tar.part005 | Other | 8.6 GB | 44968bea6b8e |
| training/libero_caddy_train.tar.part006 | Other | 8.6 GB | 6fb7a0ba3f24 |
| training/libero_caddy_train.tar.part007 | Other | 8.6 GB | ffea1906504b |
| training/libero_caddy_train.tar.part008 | Other | 8.3 GB | f8bdd88ef44e |
| training/libero_scene2_train.tar.part001 | Other | 8.6 GB | 22f74be817b7 |
| training/libero_scene2_train.tar.part002 | Other | 8.6 GB | 81298d229688 |
| training/libero_scene2_train.tar.part003 | Other | 8.6 GB | dbc314697647 |
| training/libero_scene2_train.tar.part004 | Other | 8.6 GB | b0a1afd73a0c |
| training/libero_scene2_train.tar.part005 | Other | 8.6 GB | f42973468a70 |
| training/libero_scene2_train.tar.part006 | Other | 8.6 GB | e1e05af65768 |
| training/libero_scene2_train.tar.part007 | Other | 8.6 GB | f816674d8da2 |
| training/libero_scene2_train.tar.part008 | Other | 8.6 GB | a1e304ea6a1a |
| training/libero_scene2_train.tar.part009 | Other | 8.6 GB | 03c7e4a6b917 |
| training/libero_scene2_train.tar.part010 | Other | 1.6 GB | a705eb4404c5 |
| training/robotwin_a2b_train.tar.part001 | Other | 8.6 GB | 06b67eba5b00 |
| training/robotwin_a2b_train.tar.part002 | Other | 8.6 GB | 4d9983513d43 |
| training/robotwin_a2b_train.tar.part003 | Other | 8.6 GB | 4f7487c324f8 |
| training/robotwin_a2b_train.tar.part004 | Other | 8.6 GB | 620f4f895949 |
| training/robotwin_a2b_train.tar.part005 | Other | 8.6 GB | 6e595e506d8c |
| training/robotwin_a2b_train.tar.part006 | Other | 8.6 GB | 4d3bd7edccf2 |
| training/robotwin_a2b_train.tar.part007 | Other | 8.6 GB | 12fade20790e |
| training/robotwin_a2b_train.tar.part008 | Other | 8.6 GB | 8aa0ee48fcd9 |
| training/robotwin_a2b_train.tar.part009 | Other | 1.9 GB | 81f6386a42b2 |
| .gitattributes | Repository | 2.5 KB | — |
License and Download
- License
- Not stated by the source
- Access
- No access gate
Released by Letian Cheng through its official repository on Hugging Face.