File size: 1,086 Bytes
b2b415b
74bbc85
b2b415b
 
 
 
63dd97a
b2b415b
 
 
 
 
60d4ffd
b2b415b
 
74bbc85
b2b415b
74bbc85
b2b415b
7de0174
b2b415b
 
 
 
 
 
 
 
 
 
 
 
72b6c80
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
---
title: Image Data Extractor
emoji: 🧾
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 6.20.0
python_version: "3.12"
app_file: app.py
pinned: false
models:
  - numind/NuExtract3
short_description: Extract structured JSON from documents with NuExtract3.
---

# Image Data Extractor

Image Data Extractor is a Gradio app for extracting structured JSON from document images with [NuExtract3](https://huggingface.co/numind/NuExtract3).

Upload a receipt, invoice, or other document image, optionally add accompanying text, and choose a JSON template for schema-guided structured extraction. You can also generate a grounded template from the uploaded image before running extraction.

The inference function uses Hugging Face ZeroGPU dynamic GPU allocation. Select ZeroGPU in the Space hardware settings before launching the app.

## Run locally

Install the project dependencies and launch the Hugging Face Spaces entry point:

```bash
uv sync
uv run app.py
```

The Hugging Face Space starts from `app.py`. Model loading and the ZeroGPU pipelines are defined in `model.py`.