---
license: cc-by-sa-4.0
task_categories:
  - image-classification
  - zero-shot-classification
language:
  - en
tags:
  - vision-language
  - CLIP
  - out-of-pre-training
  - OOP
  - benchmark
  - multimodal
  - few-shot
  - zero-shot
pretty_name: LAION-Beyond
size_categories:
  - 100K<n<1M

---

# LAION-Beyond: Reproducible Vision-Language Models Meet Concepts Out of Pre-Training

<p align="center">
  📄 <a href="https://openaccess.thecvf.com/content/CVPR2025/papers/Chen_Reproducible_Vision-Language_Models_Meet_Concepts_Out_of_Pre-Training_CVPR_2025_paper.pdf">Paper (CVPR 2025)</a> | 
  💻 <a href="https://github.com/M-HuangX/LAION-Beyond">Code</a> | 
  🌐 <a href="https://github.com/M-HuangX/laion_beyond">Project Page</a>
</p>

## Dataset Summary

LAION-Beyond is the **first multi-domain benchmark** specifically designed to evaluate the Out-of-Pre-training (OOP) generalization of vision-language models (e.g., CLIP, OpenCLIP, EVA-CLIP).

We distinguish two types of visual concepts:

- **IP (In-Pre-training)**: concepts that appear in the pre-training data (e.g., LAION-400M / 2B / 5B)
- **OOP (Out-of-Pre-training)**: concepts entirely absent from the pre-training data

<p align="center">
  <img src="https://raw.githubusercontent.com/M-HuangX/laion_beyond/master/static/images/Figure1_OOP_IP_difference.jpg" alt="IP vs OOP Difference" width="80%">
  <br>
  <em>Figure 1: Comparison between IP and OOP generalization. The former evaluates generalization within seen visual concepts, while the latter tests concepts absent during pre-training.</em>
</p>


The key finding of our paper is that despite OpenCLIP's image encoder forming well-separated clusters for OOP concepts, **zero-shot transfer fails significantly** due to poor image-text alignment — the token embeddings for OOP class names were never aligned with visual features during pre-training.

---

## Dataset Statistics

| Split     | Images      | Concepts |
| --------- | ----------- | -------- |
| OOP       | 106,052     | 674      |
| IP        | 51,330      | 324      |
| **Total** | **157,382** | **998**  |

<p align="center">
  <img src="https://raw.githubusercontent.com/M-HuangX/laion_beyond/master/static/images/Figure2a_LAION_Beyond_Distribution.png" width="48%">
  <img src="https://raw.githubusercontent.com/M-HuangX/laion_beyond/master/static/images/Figure2b_Image_Counts_per_category.png" width="48%">
  <br>
  <em>Figure 2: (Left) Statistics of OOP/IP concepts across different LAION scales; (Right) Detailed train/val/test split in LAION-Beyond (400M).</em>
</p>


### Domains Covered:

- 🐾 **Animals** | 🏛️ **Architecture** | 👘 **Attire**
- 🎨 **FolkArt** | 🍜 **Food** | 🦋 **Insects & Spiders**
- 🗺️ **Landmark** | 🌿 **Plants & Fungi** | 🎮 **Pokemon**

Each domain contains an IP subset and an OOP subset, covering LAION-400M, LAION-2B, and LAION-5B scales to support neural scaling law research.

---

## Dataset Structure

Each domain folder is named `{Domain}{NumClasses}_{IP/OOP}`, e.g., `Animals42_IP`, `Animals92_OOP`.

```
LAION_Beyond/
├── Animals42_IP/
│   ├── images/                        # jpg images organized by class
│   ├── label2name.json                # label index → class name
│   ├── name2label.json                # class name → label index
│   ├── merged_mapping.json            # merged label mapping
│   └── split_Xin_Animals42_IP.json    # train/val/test split info
├── Animals92_OOP/
│   └── ...
├── Architecture23_IP/
├── Architecture50_OOP/
├── Attire28_IP/
├── Attire54_OOP/
├── FolkArt27_IP/
├── FolkArt59_OOP/
├── Food27_IP/
├── Food53_OOP/
├── Insects_Spiders52_IP/
├── Insects_Spiders106_OOP/
├── Landmark30_IP/
├── Landmark59_OOP/
├── Plants_Fugi56_IP/
├── Plants_Fugi113_OOP/
├── Pokemon39_IP/
└── Pokemon89_OOP/
```

### File Descriptions

| File                  | Description                                         |
| --------------------- | --------------------------------------------------- |
| `images/`             | Raw image files (JPG), organized by class subfolder |
| `label2name.json`     | Mapping from integer label to class name string     |
| `name2label.json`     | Mapping from class name string to integer label     |
| `merged_mapping.json` | Combined label mapping across splits                |
| `split_Xin_*.json`    | Train / val / test split assignments per image      |

---

## Loading the Dataset

### Option 1: Download full dataset (recommended)

```python
from huggingface_hub import snapshot_download

local_dir = snapshot_download(
    repo_id="MHuangX/LAION-Beyond",
    repo_type="dataset",
    local_dir="./LAION_Beyond"
)
```

### Option 2: Download a single domain only

```python
from huggingface_hub import snapshot_download

local_dir = snapshot_download(
    repo_id="MHuangX/LAION-Beyond",
    repo_type="dataset",
    local_dir="./LAION_Beyond",
    allow_patterns="Animals42_IP/**" 
)
```

---

## Key Findings

1. **Strong image features for OOP concepts**: OpenCLIP's image encoder forms well-separated clusters for OOP concepts (clustering accuracy gap < 3% on most domains vs. IP concepts).

2. **Image-text alignment failure**: Zero-shot accuracy on OOP concepts is significantly lower than IP concepts, persisting even as pre-training data scales from 400M to 5B.

3. **Name-tuning is the key**: Our proposed FSNL and ZSNL algorithms, which fine-tune only the name (token) embeddings of OOP concepts, efficiently restore OOP generalization without degrading IP performance.

---

## Algorithms

### FSNL — Few-Shot Name Learning

Optimizes only OOP concept name embeddings using a few image-text pairs, with context augmentation via similar concept shuffling. Achieves state-of-the-art on 8/9 domains.

### ZSNL — Zero-Shot Name Learning

Requires no image-text pairs. Uses Novel Class Discovery (NCD) and image-text bipartite graph matching to optimize OOP name embeddings from unlabeled images only.

---

## Benchmark Results (400M split)

### OOP Few-Shot Learning (4-shot, H-mean of OOP & IP accuracy)

| Method          | Animals   | Architecture | Attire    | FolkArt   | Food     | Insects   | Landmark  | Plants    | Pokemon   | Avg       |
| --------------- | --------- | ------------ | --------- | --------- | -------- | --------- | --------- | --------- | --------- | --------- |
| OpenCLIP        | 26.75     | 30.75        | 25.88     | 35.04     | 15.36    | 22.38     | 40.25     | 21.43     | 24.48     | 26.92     |
| CoOp            | 31.37     | 57.8         | 50.39     | 52.06     | 42.55    | 25.73     | 85.89     | 24.78     | 35.52     | 45.12     |
| CLIP-Adapter    | 38.98     | 59.27        | 64.56     | 56.32     | 64.32    | 32.51     | 90.82     | 31.97     | 54.99     | 54.86     |
| **FSNL (ours)** | **46.17** | **62.63**    | **71.65** | **63.03** | **70.0** | **44.03** | **94.48** | **44.12** | **68.87** | **62.55** |

---

## Citation

If you use LAION-Beyond in your research, please cite:

```bibtex
@inproceedings{chen2025reproducible,
  title={Reproducible vision-language models meet concepts out of pre-training},
  author={Chen, Ziliang and Huang, Xin and Fan, Xiaoxuan and Wang, Keze and Zhou, Yuyu and Guan, Quanlong and Lin, Liang},
  booktitle={Proceedings of the Computer Vision and Pattern Recognition Conference},
  pages={14701--14711},
  year={2025}
}
```

---

## License

This dataset is released under the [Creative Commons Attribution-ShareAlike 4.0 International License (CC BY-SA 4.0)](http://creativecommons.org/licenses/by-sa/4.0/).

---

## Authors

[Xin Huang](https://www.linkedin.com/in/mhuangx/)†, [Ziliang Chen](https://scholar.google.com/citations?user=RC-LN4QAAAAJ&hl=en)†, Xiaoxuan Fan, [Keze Wang](https://kezewang.com/), Yuyu Zhou, [Quanlong Guan](https://scholar.google.com/citations?user=v4JiSqsAAAAJ&hl=en), [Liang Lin](http://www.linliang.net/)*

Affiliations: Peng Cheng Laboratory, Sun Yat-sen University, EPFL, Jinan University

†Equal Contribution · *Corresponding Author