Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -562,24 +562,30 @@ def main_ui():
|
|
| 562 |
with gr.Column(elem_classes=["asha-section"]):
|
| 563 |
gr.Markdown(
|
| 564 |
"""
|
| 565 |
-
|
| 566 |
-
|
| 567 |
-
|
| 568 |
-
|
| 569 |
-
|
| 570 |
-
|
| 571 |
-
|
| 572 |
-
|
| 573 |
-
|
| 574 |
-
|
| 575 |
-
|
| 576 |
-
|
| 577 |
-
|
| 578 |
-
|
| 579 |
-
|
| 580 |
-
|
| 581 |
-
|
| 582 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 583 |
""",
|
| 584 |
elem_id="asha_about"
|
| 585 |
)
|
|
|
|
| 562 |
with gr.Column(elem_classes=["asha-section"]):
|
| 563 |
gr.Markdown(
|
| 564 |
"""
|
| 565 |
+
# ASHA Form Digitizer & Hindi Voice Transcriber
|
| 566 |
+
### **Digitizing rural health, empowering ASHA workers with AI.**
|
| 567 |
+
|
| 568 |
+
This application is designed to help **ASHA (Accredited Social Health Activist) workers**—the backbone of India's rural healthcare system—quickly digitize handwritten forms and transcribe Hindi voice input. ASHA workers are often the first point of contact for healthcare in villages, but their work is slowed by manual paperwork and language barriers. This tool streamlines their workflow, making data entry faster, more accurate, and accessible even for those more comfortable with Hindi speech than typing.
|
| 569 |
+
|
| 570 |
+
- **Image-based field extraction:** Upload a photo of an ASHA form and the app will automatically detect and extract all field labels, ready for digital entry.
|
| 571 |
+
- **Hindi voice transcription:** Fill any field by speaking in Hindi (Devanagari script) for instant, accurate transcription.
|
| 572 |
+
- **Data export:** All submitted data is saved in a CSV for further use or analysis.
|
| 573 |
+
|
| 574 |
+
---
|
| 575 |
+
|
| 576 |
+
**Powered by Gemma 3n 4B**
|
| 577 |
+
This demo runs on the **Gemma 3n 4B model**, a lightweight yet capable model for on-device AI tasks like image-to-text and Hindi speech transcription.
|
| 578 |
+
|
| 579 |
+
**On-device CPU Loading**
|
| 580 |
+
Since this application runs entirely **on-device using CPU**, the model takes some time to **load the first time**. After the initial load, processing is smooth and does not require an internet connection.
|
| 581 |
+
|
| 582 |
+
---
|
| 583 |
+
|
| 584 |
+
**Why this matters:**
|
| 585 |
+
ASHA workers serve over 900 million people in rural India, often with limited digital literacy and resources. By making form digitization and voice transcription seamless, this app saves time, reduces errors, and helps bring rural health data into the digital age—empowering both workers and the communities they serve.
|
| 586 |
+
|
| 587 |
+
**Note:**
|
| 588 |
+
While you are reading this, the **Gemma 3n model** is being loaded in the background to ensure a smooth and fast demo experience. Please explore each step—the workflow is strictly gated for demo clarity. All features are designed for real-world usability and hackathon evaluation.
|
| 589 |
""",
|
| 590 |
elem_id="asha_about"
|
| 591 |
)
|