---
language:
- en
tags:
- arxiv:2609.24749
- world-model
- jepa
- robotics
- decision-alignment
library_name: pytorch
datasets:
- Shuaijun/D-JEPA-Dataset
---

# D-JEPA checkpoints

**[Project website](https://nebulis-lab.com/D-JEPA)** ·
**[Paper](https://arxiv.org/abs/2609.24749)** ·
**[Model repository](https://huggingface.co/Shuaijun/D-JEPA)** ·
**[Decision-supervision dataset](https://huggingface.co/datasets/Shuaijun/D-JEPA-Dataset)** ·
**[Code](https://github.com/NEBULIS-Lab/D-JEPA)**

Task/module profiles for **D-JEPA: A Decision-Aligned Latent World Model**.
See each profile's `README.md` and `config.json` for architecture, provenance,
upstream dependencies, calibrated decisions and tensor-file checksums.

## Choose a checkpoint profile

| Task | Profile | Contents |
|---|---|---|
| PushT | [Relational alignment](./pusht-relational/README.md) | Learned decision-alignment module |
| PushT | [Predictive adaptation](./pusht-predictive-adaptation/README.md) | Updated parameters; requires the exact base model |
| PushT | [Adapted predictor](./pusht-adapted-predictor-full/README.md) | Full predictor weights |
| PushT | [Multi-geometry alignment](./pusht-multi-geometry/README.md) | Module using four predictive geometries |
| PushT | [Exact representation realization](./pusht-exact-realization-full/README.md) | Two predictors and relation module in one tensor file |
| Reacher | [Relational alignment](./reacher-relational/README.md) | Task-specific decision-alignment module |
| Reacher | [Temporal transport](./reacher-temporal-transport/README.md) | Bounded future-update module |
| Reacher | [Full model](./reacher-world-model-full/README.md) | Predictors, alignment and transport weights |
| Granular | [Relational alignment](./granular-relational/README.md) | Spatial-feature alignment module |
| Granular | [Multi-view alignment](./granular-multiview/README.md) | Four-view variant using the same backbone |
| PushObj | [Unseen-shape transfer](./pushobj-unseen-shapes/README.md) | Task-local ordinal alignment module |
| PushT | [Visual-shift alignment](./pusht-visual-shifts/README.md) | Corruption-trained ordinal alignment module |

These are configurations of **D-JEPA**, not 12 separate methods. Each profile
contains `model.pt`, `config.json` and its own card. The [manifest](./manifest.json)
indexes source and exported weight hashes.

## Loading and dependencies

Download an individual checkpoint profile with the Hugging Face CLI:

```bash
hf download Shuaijun/D-JEPA \
  pusht-relational/model.pt pusht-relational/config.json pusht-relational/README.md \
  --local-dir checkpoints/D-JEPA
```

This repository hosts custom PyTorch checkpoint profiles, not dataset splits.
It does not require the Dataset Viewer. The linked supervision dataset is a
downloadable ZIP/NPZ archive; see its card for extraction and loading instructions.

All `model.pt` files contain tensor dictionaries and support
`torch.load(path, map_location="cpu", weights_only=True)`. Exact tensor values
from selected formal checkpoints are preserved. Packaging removes optimizer
state and private path metadata; exported file hashes are therefore distinct
from original training-file hashes.

```python
import torch

state = torch.load("checkpoints/D-JEPA/pusht-relational/model.pt",
                   map_location="cpu", weights_only=True)
```

This reads a tensor state dictionary, not an instantiated model. Use the matching
D-JEPA architecture and configuration; these custom modules are not loaded through
Transformers `AutoModel.from_pretrained`.

Profiles with `upstream_included: true` contain pretrained predictors. Other
profiles contain our learned modules or updated parameter subsets and require
their recorded predictive models/features. `pusht-exact-realization-full` is
one file containing two predictors and a relational operator, not a distilled
single-backbone student.

Reacher's physical result uses the relational selector; temporal transport is
also provided but its native-distance diagnostic is not a substitute result.
Multi-geometry's 128-start development result remains a separate protocol from
the independent 256-start PushT evaluation.

Local release candidate: authors have not yet selected publication licenses or
completed all upstream weight redistribution checks. This card grants no new
rights to third-party material. The author-designated repository is
`Shuaijun/D-JEPA`.