|
|
--- |
|
|
title: Gemma‑3n SOAP Generator |
|
|
emoji: 🩺 |
|
|
colorFrom: blue |
|
|
colorTo: green |
|
|
sdk: gradio |
|
|
sdk_version: 5.38.2 |
|
|
app_file: app.py |
|
|
pinned: false |
|
|
--- |
|
|
|
|
|
# Gemma‑3n SOAP Generator Space |
|
|
|
|
|
This Space uses the `google/gemma-3n-e2b-it` model to: |
|
|
|
|
|
1. Generate 100 synthetic doctor’s notes and corresponding SOAP “ground truths” on startup. |
|
|
2. Save `doc_notes.tsv` and `ground_truth_soap.tsv` to the workspace. |
|
|
3. Expose a no‑op Gradio UI stub for demonstration. |
|
|
|
|
|
## SDK & Dependencies |
|
|
|
|
|
Make sure you have the following installed (see `requirements.txt`): |
|
|
|
|
|
- **Transformers** (from source): `git+https://github.com/huggingface/transformers.git` |
|
|
- **BitsAndBytes** for 8‑bit quantization |
|
|
- **Accelerate** for device mapping |
|
|
- **Timm** (model helpers) |
|
|
- **EasyOCR** (if using OCR features) |
|
|
- **Gradio** for the UI |
|
|
- **Hugging Face Hub SDK**: `huggingface_hub` |
|
|
|
|
|
Install them all via: |
|
|
|
|
|
```bash |
|
|
pip install --no-deps \ |
|
|
git+https://github.com/huggingface/transformers.git \ |
|
|
bitsandbytes accelerate timm easyocr gradio huggingface_hub |
