---
dataset_info:
  features:
  - name: image_url
    dtype: string
  - name: show
    dtype: string
  - name: designer
    dtype: string
  - name: embedding
    list: float32
    length: 768
  splits:
  - name: train
    num_bytes: 3009108815
    num_examples: 933328
  download_size: 3463198535
  dataset_size: 3009108815
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---
# Vogue Runway Image-Embedding Corpus

Dataset of [Vogue Runway](https://www.vogue.com/fashion-shows) images and embeddings.

| Metric | Value |
|--------|-------|
| **Designers** | 1 749 |
| **Collections** | 25 876 |
| **Images / Looks** | 933 328 |
| **Years covered** | 1990 → 2025 |
| **Embedding model** | `google/vit-base-patch16-224` (mean-pooled CLS tokens) |
| **Vector size** | 768 float32 |


```python
from datasets import load_dataset
ds = load_dataset("tonyassi/vogue933k-embedding", streaming=True, split="train")
for i, row in zip(range(10), ds): print(row)
```