README.md

---
title: README
emoji: 🧥
colorFrom: indigo
colorTo: green
sdk: static
pinned: false
license: mit
short_description: Open-source fashion intelligence and visual search.
---

<section>
  <h1>HopitAI</h1>
  <p>
    HopitAI builds open-source fashion intelligence models for visual search,
    product matching, catalog retrieval, and ecommerce discovery.
  </p>
</section>

<section>
  <h2>MODA Fashion Retrieval</h2>
  <p>
    MODA is a family of fashion retrieval models covering both halves of the
    problem: finding products from an image, and finding products from a text
    description. Every model and result is MIT licensed and reproducible from
    the public repository.
  </p>
  <p>
    <a href="https://huggingface.co/spaces/HopitAI/moda-fashion-search">Try the interactive demo</a>
    ·
    <a href="https://github.com/hopit-ai/Moda">View the GitHub repo</a>
    ·
    <a href="https://huggingface.co/HopitAI/moda-fashion-deepfashion2">Start with DeepFashion2</a>
  </p>
</section>

<section>
  <h2>Featured Models</h2>
  <table>
    <tr>
      <th>Model</th>
      <th>Task</th>
      <th>Best for</th>
    </tr>
    <tr>
      <td><a href="https://huggingface.co/HopitAI/moda-fashionsiglip-multiview-203m">MODA-FashionSigLIP-MultiView-203M</a></td>
      <td>Text to image</td>
      <td>Searching a catalog in plain language. Multi-view retrieval system over the frozen FashionSigLIP checkpoint, adding zero parameters.</td>
    </tr>
    <tr>
      <td><a href="https://huggingface.co/HopitAI/moda-fashion-distilled">MODA-Fashion-Distilled</a></td>
      <td>Image to image</td>
      <td>Best overall quality</td>
    </tr>
    <tr>
      <td><a href="https://huggingface.co/HopitAI/moda-fashion-distilled-512d">MODA-Fashion-Distilled-512d</a></td>
      <td>Image to image</td>
      <td>Compact vector search indexes</td>
    </tr>
    <tr>
      <td><a href="https://huggingface.co/HopitAI/moda-fashion-deepfashion2">MODA-Fashion-DeepFashion2</a></td>
      <td>Image to image</td>
      <td>Simple reproducible recipe</td>
    </tr>
    <tr>
      <td><a href="https://huggingface.co/HopitAI/moda-fashion-vision-fp16">MODA-Fashion-Vision-FP16</a></td>
      <td>Image to image</td>
      <td>Smaller edge and mobile deployment</td>
    </tr>
    <tr>
      <td><a href="https://huggingface.co/HopitAI/moda-fashion-matryoshka">MODA-Fashion-Matryoshka</a></td>
      <td>Image to image</td>
      <td>Flexible embedding dimensions</td>
    </tr>
  </table>
</section>

<section>
  <h2>Quick Start</h2>

Image to image, for finding visually similar products:

```bash
huggingface-cli download HopitAI/moda-fashion-deepfashion2 --local-dir ./moda-fashion-deepfashion2
```

```python
import open_clip

model, preprocess_train, preprocess_val = open_clip.create_model_and_transforms(
    "hf-hub:HopitAI/moda-fashion-deepfashion2"
)
```

Text to image, for searching a catalog in plain language:

```bash
pip install "git+https://huggingface.co/HopitAI/moda-fashionsiglip-multiview-203m"
```

```python
from moda_fashionsiglip_multiview import ModaFashionSigLIP

retriever = ModaFashionSigLIP.from_pretrained()
index = retriever.build_index(image_paths, item_ids=item_ids)
results = retriever.search("red floral summer dress", index, top_k=5)[0]
```
</section>

<section>
  <h2>Use Cases</h2>
  <ul>
    <li>Search a fashion catalog with a natural-language query.</li>
    <li>Find visually similar products in a fashion catalog.</li>
    <li>Match street-style looks to shoppable items.</li>
    <li>Deduplicate product images across marketplaces.</li>
    <li>Build embedding indexes for ecommerce search and recommendations.</li>
  </ul>
</section>