Text Generation
PEFT
Safetensors
English
emergency-response
resqai
structured-json
lora
qwen2.5
Indian-emergency-system
Instructions to use Devrcb/Emergency_informed_Qwen with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use Devrcb/Emergency_informed_Qwen with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
readme added
Browse files
README.md
CHANGED
|
@@ -1,3 +1,55 @@
|
|
| 1 |
---
|
| 2 |
-
license:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: Qwen/Qwen2.5-0.5B-Instruct
|
| 4 |
+
library_name: peft
|
| 5 |
+
pipeline_tag: text-generation
|
| 6 |
+
language:
|
| 7 |
+
- en
|
| 8 |
+
tags:
|
| 9 |
+
- emergency-response
|
| 10 |
+
- resqai
|
| 11 |
+
- structured-json
|
| 12 |
+
- lora
|
| 13 |
+
- qwen2.5
|
| 14 |
+
- Indian-emergency-system
|
| 15 |
---
|
| 16 |
+
|
| 17 |
+
# Devrcb/Emergency_informed_Qwen
|
| 18 |
+
|
| 19 |
+
**Emergency_informed_Qwen** is a fine-tuned LoRA adapter based on [`Qwen/Qwen2.5-0.5B-Instruct`](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct)[cite: 2]. It is specifically trained as the **Emergency Understanding Agent** for the **ResQAI Emergency Response System**[cite: 2, 3].
|
| 20 |
+
|
| 21 |
+
The model parses unstructured emergency reports (voice transcripts, text messages, citizen distress calls) and converts them into strict, structured JSON outputs for automated emergency dispatch, triage, and location routing[cite: 2, 3].
|
| 22 |
+
|
| 23 |
+
---
|
| 24 |
+
|
| 25 |
+
## Model Details
|
| 26 |
+
|
| 27 |
+
* **Developed by:** Devrcb
|
| 28 |
+
* **Model Type:** Causal Language Model (Fine-Tuned Adapter via PEFT/LoRA)
|
| 29 |
+
* **Base Model:** `Qwen/Qwen2.5-0.5B-Instruct`
|
| 30 |
+
* **Task:** Zero-shot / Few-shot Emergency Classification & Structured JSON Extraction
|
| 31 |
+
* **Fine-Tuning Framework:** Hugging Face `trl` (SFTTrainer) & `peft`
|
| 32 |
+
* **Language:** English (optimized for Indian emergency context)
|
| 33 |
+
|
| 34 |
+
---
|
| 35 |
+
|
| 36 |
+
## Intended Use
|
| 37 |
+
|
| 38 |
+
### Primary Use Case
|
| 39 |
+
This model is designed to run in real-time emergency dispatch pipelines (e.g., ResQAI)[cite: 2, 3]. Given a raw description of an emergency, it extracts:
|
| 40 |
+
1. **Emergency Type & Severity Level**
|
| 41 |
+
2. **Services Required** (EMS, Fire Brigade, Police)[cite: 2]
|
| 42 |
+
3. **Medical Speciality Needed**[cite: 2]
|
| 43 |
+
4. **Immediate Citizen First-Aid / Action Advice**[cite: 2]
|
| 44 |
+
|
| 45 |
+
### Output Format
|
| 46 |
+
The model is strictly trained to output **only valid JSON** adhering to the following schema[cite: 2]:
|
| 47 |
+
|
| 48 |
+
```json
|
| 49 |
+
{
|
| 50 |
+
"citizen_advice": ["Call emergency services immediately", "Stay clear of smoke"],
|
| 51 |
+
"emergency_type": "Fire",
|
| 52 |
+
"medical_speciality": "Emergency Medicine",
|
| 53 |
+
"services_required": ["Emergency Medical Services (EMS)", "Firefighting"],
|
| 54 |
+
"severity": "High"
|
| 55 |
+
}
|