# Jin — Hanzo AI

> Open research on Joint-Embedding Predictive Architectures (JEPA): self-supervised visual representations, world models, and multimodal learning.

hanzoai/jin

# Jin

Research on learning without labels

A joint-embedding predictive architecture does not reconstruct pixels. It predicts the embedding of a piece of an image from the embedding of another piece — so the model spends its capacity on what the image means rather than on what it looks like. Jin is our work on that idea, in the open, with the code and the papers to read.

[Read the docs](https://docs.hanzo.ai)[View on GitHub](https://github.com/hanzoai)

What is in the repository

- I-JEPA01
- Saccade JEPA02
- Three losses at once03
- Masked autoencoding04
- Ways to see what it learned05
- The papers06

A clean implementation of the image variant: a context encoder, a target encoder that is its moving average, and a predictor that has to guess the target embeddings. Both a ViT and an energy-transformer backbone.

Ours. Eyes jitter constantly and the brain predicts what the jitter will do to the scene, so the model shifts an image, predicts how its embedding moves, and shifts back to check it lands where it started.

Huber loss on the prediction, a cycle-consistency term that punishes drift on the round trip, and the invariance and covariance parts of VICReg to stop the representation collapsing.

An MAE with self-distillation sits alongside, for the comparison that matters: is a decoder worth its parameters when the goal is a good representation rather than a good reconstruction?

Self-supervised training has no validation loss to watch, so there are linear probes, KNN, attention-map visualisation in a Dash dashboard, correlation dimension, and UMAP coloured by true class.

Grant proposals on hierarchical JEPA for edge AI, where the interesting constraint is learning useful representations on hardware that cannot phone home.

Paste and ship

Paste this into any agent. It reads the skill manifest and calls Jin from there.

Agent promptCopy

```
Read https://hanzo.ai/skill.md and use Hanzo Jin in my project.
```

I-JEPAReference build

SaccadeOur own variant

MAESelf-distilled

PapersEdge AI proposals

## Read the code

[Read the docs](https://docs.hanzo.ai)[View on GitHub](https://github.com/hanzoai)

## Get Jin

Multimodal foundation model

License: Apache-2.0[hanzoai](https://github.com/hanzoai)

[Deploy to Cloud](https://platform.hanzo.ai/deploy)