File size: 919 Bytes
97985eb
 
 
 
 
 
 
 
 
 
 
 
 
9b9c8e7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
97985eb
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: MedScribe AI
emoji: 📊
colorFrom: yellow
colorTo: red
sdk: gradio
sdk_version: 5.49.1
app_file: app.py
pinned: false
license: mit
short_description: This Space demonstrates a Transformer-based clinical Named E
---

# Clinical NER with Text

This Space demonstrates a Transformer-based clinical Named Entity Recognition (NER) application with a bonus multimodal audio → text pipeline.

- **Text tab:** Extracts biomedical entities from user-provided clinical text.
- **Audio tab:** Transcribes short speech using Whisper and runs the same NER on the transcript.

> **Disclaimer**: Research/education demo only. Not a medical device and not for clinical use.

## Models
- NER: `d4data/biomedical-ner-all`
- ASR: `openai/whisper-tiny.en`

## Run locally
```bash
pip install -r requirements.txt
python app.py


Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference