# Jeveration: System One Models Now Generate

> Jeveration generates coherent 512x512 images in a single 78ms forward pass by conditioning parallel code predictions on a shared random state, achieving 122x speedup over Janus-Pro-7B.

- **Source:** [arXiv](https://arxiv.org/abs/2611.24738)
- **Published:** 2026-11-20
- **Permalink:** https://picx.dev/p/SRsVgf
- **Whiteboard:** https://picx.dev/p/SRsVgf/image

## Summary

# Jeveration: System One Models Now Generate

## Summary (Overview)

- **New generation paradigm**: Jeveration introduces a third paradigm for image generation—parallel typed decisions in a single forward pass—contrasting with autoregressive and diffusion approaches that require sequential sampling.
- **Shared random state for coherence**: The key innovation is conditioning all code positions on a shared random state ξ, enabling parallel predictions to converge on globally coherent images despite the many-to-one prompt-to-image mapping.
- **Two-phase training**: The method first learns the image distribution via masked prediction, then uses distribution matching (Di[M]O-style) with confidence-guided correction to teach the model to use the random state effectively.
- **Unified architecture**: A single frozen VLM (Qwen3.5-4B) with shared LoRA adapters handles text-to-image, image-to-image, and dense prediction tasks in one checkpoint.
- **Exceptional efficiency**: Generates 512×512 images in 78 ms on one H200, achieving 122× speedup over Janus-Pro-7B while maintaining competitive quality (0.74 GenEval).

## Introduction and Theoretical Foundation

### Background and Motivation

Image generation is dominated by two paradigms:
- **Autoregressive models** [11, 40, 44]: generate image tokens sequentially
- **Diffusion models** [8, 25, 48]: iterative denoising

Both require many sequential model evaluations, increasing inference latency. Jeveration introduces a third paradigm built on **System One models** (e.g., Jev [1]) that return typed decisions (Choices, Scores) in a single parallel query without extended reasoning.

### The Core Challenge: Global Consistency

A single prompt admits many valid images, yet parallel predictions must converge on a coherent whole. The theoretical foundation rests on a key information-theoretic insight:

For independent parallel predictions, the best forward-KL approximation to the true target distribution $p(c|y)$ satisfies:

$$\min_{q_{\text{ind}}} \mathrm{KL}(p \| q_{\text{ind}}) = \sum_{i=1}^{N} H(c_i \mid y) - H(c \mid y)$$

where $H$ denotes entropy. The gap is the target's **total correlation**—the measure of how strongly codes depend on each other. A coherent image requires this dependence, but independent sampling matches only the marginals $p(c_i|y)$, producing incompatible image regions.

### Solution: Shared Random State

Jeveration conditions every code position on a shared random state sequence ξ:

$$q_{\theta}(c \mid y) = \mathbb{E}_{\xi \sim p_0} \left[ \prod_{i=1}^{N} q_{\theta,i}(c_i \mid y, \xi) \right]$$

Although codes are sampled independently given $(y, \xi)$, their shared dependence on ξ allows the image distribution to capture inter-position dependencies.

## Methodology

### Architecture

The system builds on a **frozen Qwen3.5-4B VLM** with:
- **Shared LoRA adapters** (rank-256) across all tasks
- **Task-specific output heads**
- **TA-TiTok image tokenizer** with N = 128 codes per image

**Decision tokens** are appended to the input:
- *Image generation*: Initialized from sampled random state ξ (masking with probability ρ = 0.6, uniform random codes for unmasked positions, Gaussian noise σ = 0.1)
- *Dense prediction*: Serve as spatial queries aligned with visual tokens

### Image Code Prediction

Each decision token's final hidden state $u_i$ feeds two complementary branches:

$$q_{\theta,i}(\cdot \mid y, \xi) = \operatorname{softmax}\left(f_{\mathrm{cb}}(u_i) + f_{\text{wide}}(u_i)\right)$$

- **Codebook branch** $f_{\mathrm{cb}}$: projects to code space with classifier initialized from tokenizer's codebook
- **Wide branch** $f_{\mathrm{wide}}$: zero-initialized direct mapping to code logits

### Two-Phase Training

**Phase 1: Learning the image distribution (masked prediction)**

$$\mathcal{L}_{\mathrm{mask}}(\phi) = \mathbb{E} \left[ \frac{1}{|M|} \sum_{i \in M} - \log p_{\phi,i}(c_i \mid c_{\bar{M}}, y) \right]$$

The visible codes supply image-specific context, enabling the model to learn inter-position dependencies.

**Phase 2: Learning image decisions (distribution matching)**

$$\mathcal{L}_{\mathrm{DM}}(\theta) = \mathbb{E} \left[ \frac{1}{|M|} \sum_{i \in M} \mathrm{sg}(p_{\psi,i} - p_{\phi,i})^{\top} \ell_{\theta,i} \right]$$

where:
- $\phi$ = frozen first-phase model (represents image distribution)
- $\psi$ = auxiliary model tracking θ's current distribution
- $\ell_{\theta,i}$ = logits from the decision model
- sg = stop-gradient

**Confidence-guided correction**: Re-mask the half of positions with lowest confidence $q_{\theta,i}(\hat{c}_i|y,\xi)$, complete via frozen model, then train θ to predict corrected codes:

$$\mathcal{L}_{\mathrm{gen}} = \mathcal{L}_{\mathrm{DM}} + \lambda \mathcal{L}_{\mathrm{corr}}$$

## Empirical Validation / Results

### Text-to-Image Generation

**Key results (Table 1)**:

| Model | Unified | Params (B) | Steps | Res. (px) | Speedup↑ | GenEval↑ | Latency (ms)↓ |
|-------|---------|------------|-------|-----------|----------|----------|---------------|
| Janus-Pro-7B | √ | 7.0 | 576 | 384 | 1.0× | 0.80 | 9,508.4 |
| Show-o Turbo-512 | √ | 1.3 | 4 | 512 | 101.0× | 0.63 | 94.1 |
| **Jeveration** | **√** | **5.2** | **1** | **512** | **122.2×** | **0.74** | **77.8** |

- **122× faster** than Janus-Pro-7B
- **0.11 GenEval improvement** over Show-o Turbo-512 at same resolution
- Higher GenEval than SANA-Sprint 0.6B, FLUX.1-schnell, and SDXL-DMD2 (one-step 1024-px models)

### Image-to-Image Generation (Table 2)

| Model | HED | Depth | Colorize | Style PSNR↑ |
|-------|-----|-------|----------|-------------|
| **Jeveration** | **0.880** | **0.802** | **0.937** | **15.40** |
| Copy the input | 0.828 | 0.713 | 0.998 | - |
| Decoder without source | 0.792 | 0.767 | 0.757 | 14.37 |

Strong structural fidelity across edge maps, depth maps, colorization, and style transfer (anime, Studio Ghibli, watercolor).

### Dense Prediction on NYUv2 (Table 3)

| Model | Seg. mIoU↑ | Depth abs.↓ | Depth δ₁↑ | Normal mean (°)↓ |
|-------|-----------|-------------|-----------|-------------------|
| **Jeveration** | **67.7** | 0.267 | 0.898 | 24.2 |
| Dense-only (one task) | 67.8 | **0.246** | **0.915** | **21.2** |
| DeepLabV3+ R50 | 53.5 | 0.393 | - | 22.0 |

Jeveration outperforms DeepLabV3+ by **14.2 mIoU** and reduces depth error by **32%**.

### Ablation Studies (Table 4)

| Variant | GenEval↑ | HPS↑ |
|---------|----------|------|
| **w/ random state** | **0.73** | **28.27** |
| w/ fixed state | 0.70 | 26.63 |
| **w/ correction target** | **0.74** | **28.03** |
| w/o correction target | 0.73 | 27.92 |
| **Image head (wide + codebook)** | **0.64** | **25.42** |
| Image head (wide only) | 0.60 | 25.34 |
| Image head (codebook only) | 0.59 | 25.04 |

Key findings:
- Random state improves coherence: +0.03 GenEval, +1.64 HPS over fixed state
- Correction target refines details (HPS +0.11)
- Combined head branches outperform either alone (complementary benefits)

### Joint Training (Table 5)

Training on all tasks costs almost no text-to-image quality (GenEval 0.74 vs 0.73; HPS 28.06 vs 28.28), demonstrating minimal task interference.

## Theoretical and Practical Implications

### Theoretical Contributions

1. **Extends System One paradigm**: Demonstrates that decision models can generate, not just decide—broadening the theoretical scope of parallel typed decision frameworks.
2. **Information-theoretic framing**: Formally characterizes why independent parallel prediction fails (total correlation gap) and how shared randomness bridges this gap.
3. **Distribution matching for parallel generation**: Shows that masked models can serve as training targets for single-pass generators within a frozen backbone, avoiding the need to train generative models from scratch.

### Practical Implications

1. **Dramatic latency reduction**: 78 ms per image enables real-time and edge deployment scenarios previously infeasible with autoregressive or diffusion approaches.
2. **Unified visual intelligence**: One checkpoint serving generation, editing, and perception tasks simplifies deployment and reduces memory footprints.
3. **Frozen backbone reuse**: Leveraging pretrained VLMs with lightweight adapters makes the approach accessible—training takes only ~4 days on 4× RTX PRO 6000 GPUs.
4. **Controllable generation at near-zero cost**: Image-to-image tasks add only ~14 ms over text-to-image, making conditional generation practical for interactive applications.

## Conclusion

Jeveration establishes a third paradigm for image generation where System One models return complete images as single joint decisions. The key innovations—shared random state conditioning and two-phase distribution-matching training—enable globally coherent images from parallel predictions.

**Main takeaways**:
- A frozen VLM can generate high-quality images (0.74 GenEval) in a single forward pass (78 ms)
- One checkpoint unifies text-to-image, image-to-image, and dense prediction with minimal task interference
- The paradigm extends naturally to diverse visual tasks

**Future directions**:
- Extension to further visual tasks and modalities (video, 3D, audio)
- Scaling to higher resolutions and larger backbones
- Exploring additional applications of the shared random state mechanism
- Potential integration with other System One decision interfaces (e.g., Box, Scores) for richer multimodal interaction

---

_Markdown view of https://picx.dev/p/SRsVgf, served by PicX — AI-generated visual whiteboard summaries of research papers._
