Downloads · 30 days
0
cjepa-pusht-videosaur-planning
Ryukijano/cjepa-pusht-videosaur-planning
cjepa-pusht-videosaur-planning is a reinforcement learning model from Ryukijano. Use it for the reinforcement learning task on the model card, and read the license before you ship it in a product. It is set up for pytorch. The card lists the license as apache-2.0.
Lean public mirror of the three Push-T VideoSAUR policies used in the ICML 2026 agent reproduction of Causal-JEPA (Claim 2 planning matrix).
Access
Public
Updated Aug 2, 2026
Repo size
317 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.ckpt317 MB · 100%
At a glance
- Task
- Reinforcement Learning
- Library
- pytorch
- License
- apache-2.0
- Access
- Public
- Created
- Aug 2, 2026
- Updated
- Aug 2, 2026
- SHA
- ec53154b
Model card
From the Hugging Face model README
C-JEPA Push-T VideoSAUR planning checkpoints
Lean public mirror of the three Push-T VideoSAUR policies used in the ICML 2026 agent reproduction of Causal-JEPA (Claim 2 planning matrix).
| File | Mask |M| | Role in repro |
|---|---|---|
pusht_videosaur_0_epoch_30_object.ckpt | 0 | OC-JEPA baseline policy |
pusht_videosaur_1_epoch_30_object.ckpt | 1 | Primary C-JEPA setting (paper Table 6) |
pusht_videosaur_2_epoch_30_object.ckpt | 2 | C-JEPA |M|=2 |
Each file is ~101 MB. This repo does not include multi-GB slot/rollout pickles.
Links
- Paper (OpenReview): https://openreview.net/forum?id=VMAHQDOtjp
- arXiv: https://arxiv.org/abs/2602.11389
- Reproduction logbook: https://huggingface.co/spaces/Ryukijano/repro-causal-jepa-learning-world-models-through-object-level-latent-masking
- Upstream full checkpoint dump: https://huggingface.co/HazelNam/CJEPA
- Code: https://github.com/galilai-group/cjepa
License
Apache-2.0, matching HazelNam/CJEPA.
Provenance
Copied from the official cjepa-ckpts/ files used by this reproduction (same bytes as HazelNam/CJEPA). Planning was run with stable-worldmodel CEM under history_keys=('pixels','proprio').
- Task
- Reinforcement Learning
- Library
- pytorch
- License
- apache-2.0
- Created
- Aug 2, 2026
- Updated
- Aug 2, 2026