---
library_name: silva
pipeline_tag: image-classification
license: mit
base_model: google/siglip2-so400m-patch14-384
base_model_relation: adapter
tags:
  - aesthetic
  - siglip2
  - ordinal-regression
metrics:
  - spearmanr
  - pearsonr
  - mae
model-index:
  - name: silva-luna
    results:
      - task:
          type: image-classification
          name: Distilled Aesthetic Scoring
        dataset:
          type: Jannchie/silva-luna-25k
          name: silva-luna-25k
        metrics:
          - type: spearmanr
            value: 0.8105
          - type: pearsonr
            value: 0.7986
          - type: mae
            value: 0.4331
---

<p align="center">
  <img src="https://raw.githubusercontent.com/Jannchie/silva/main/assets/silva-header.png" alt="SILVA" width="100%">
</p>

# SILVA — Distilled VLM Judge

**[▶ Try it in your browser](https://huggingface.co/spaces/Jannchie/silva-aesthetic-demo)** — upload an illustration, see this judge's score live.

Scores an illustration by the taste of a **VLM judge**, distilled — not a universal quality
model, so it won't match anyone else's preferences. Output is a single number in
`[0, 1]`; higher means more to this judge's liking.

**Only the head ships here (~7 MB), not an image model.** It runs on top of the frozen
`google/siglip2-so400m-patch14-384` backbone, which `silva[backbone]` installs and loads for you.

## Quickstart

```python
# pip install "silva-scorer[backbone] @ git+https://github.com/Jannchie/silva"
from silva import SilvaScorer

scorer = SilvaScorer.from_pretrained("Jannchie/silva-luna")
print(scorer.score("your_image.jpg"))     # 0.73
print(scorer.score(["a.jpg", "b.jpg"]))    # [0.73, 0.41]
```

Already have `google/siglip2-so400m-patch14-384` embeddings? Skip the backbone and score them directly:

```python
# pip install "silva-scorer @ git+https://github.com/Jannchie/silva"
from silva import EmbeddingAestheticModel

head = EmbeddingAestheticModel.from_pretrained("Jannchie/silva-luna").eval()
score = head(embedding)["calibrated_score"]  # calibrated to the label distribution; ["score"] for raw. embedding: [B, 1152] pooler_output
```

## Scores (held-out test split)

| Spearman | Pearson | MAE (1–5) | Top-5% |
|---|---|---|---|
| 0.8105 | 0.7986 | 0.4331 | 0.4672 |

Architecture: `embedding[1152] → LayerNorm → MLP [1024, 512, 256] → ordinal head`. Trained on
rankings from `openai:gpt-5.6-luna`, which ordered eight illustrations at a time; the orderings were pooled with Plackett-Luce into one latent per picture (degree 14, split-half reliability 0.827). The labels ARE published -- see the dataset link above. [Source](https://github.com/Jannchie/silva)

## Citation

```bibtex
@software{pan2026silva,
  author  = {Pan, Jianqi},
  title   = {{SILVA}: {SigLIP}-based Illustration Visual Aesthetic Scorer},
  year    = {2026},
  url     = {https://github.com/Jannchie/silva},
}
```