|
|
--- |
|
|
library_name: transformers |
|
|
pipeline_tag: text-generation |
|
|
tags: |
|
|
- medical |
|
|
- chatbot |
|
|
- instruction-tuning |
|
|
- LLaMA |
|
|
- LoRA |
|
|
base_model: meta-llama/Llama-3.2-3B-Instruct |
|
|
datasets: |
|
|
- ruslanmv/ai-medical-chatbot |
|
|
language: en |
|
|
license: apache-2.0 |
|
|
--- |
|
|
|
|
|
# Hippocratically LLaMA 3.2: Medical Chatbot (LoRA Fine-tuned) |
|
|
|
|
|
**Model:** thebnbrkr/hippocratically_llama3.2 |
|
|
**Base:** `meta-llama/Llama-3.2-3B-Instruct` |
|
|
**Fine-tuning:** LoRA (via PEFT + TRL’s SFTTrainer) |
|
|
**Dataset:** `ruslanmv/ai-medical-chatbot` (public medical Q&A) |
|
|
|
|
|
--- |
|
|
|
|
|
## Model Description |
|
|
|
|
|
`hippocratically_llama3.2` is a LoRA-finetuned version of Meta’s LLaMA 3.2B-Instruct model, adapted for clinical dialogue and general medical question-answering. It uses instruction-tuning and parameter-efficient fine-tuning (PEFT) methods to create a compact and responsive healthcare assistant for research and educational use. |
|
|
|
|
|
--- |
|
|
|
|
|
## Intended Use |
|
|
|
|
|
- Medical Q&A chatbot simulations |
|
|
- Research on LoRA-based instruction tuning |
|
|
- Prompt evaluation in clinical contexts |
|
|
- LLM alignment/hallucination analysis |
|
|
|
|
|
--- |
|
|
|
|
|
## Limitations |
|
|
|
|
|
- Not suitable for clinical deployment |
|
|
- May hallucinate outdated or unsafe information |
|
|
- Doesn't prioritize emergencies unless prompted |
|
|
- Lacks real-time patient data or diagnostic context |
|
|
|
|
|
--- |
|
|
|
|
|
## Training Details |
|
|
|
|
|
- **Base model:** `meta-llama/Llama-3.2-3B-Instruct` |
|
|
- **Fine-tuning method:** LoRA (`r=8`, `alpha=16`, dropout `0.05`) |
|
|
- **Target modules:** `q_proj`, `v_proj` |
|
|
- **Trainer:** `SFTTrainer` from `trl` |
|
|
- **Batch size:** 2 |
|
|
- **Epochs:** 1 |
|
|
- **Learning rate:** 2e-4 |
|
|
- **Precision:** fp16 |
|
|
- **Output dir:** `./llama3.2-medical-chat` |
|
|
|
|
|
--- |
|
|
|
|
|
## Example Prompt: |
|
|
prompt: | |
|
|
How can someone determine their metabolic rate? |
|
|
|
|
|
## Example response: |
|
|
Metabolic rate can be determined in several ways, depending on the context and level of precision required. |
|
|
|
|
|
1. **Basal Metabolic Rate (BMR):** This is the number of calories your body needs at rest. A commonly used formula to estimate BMR for women is: |
|
|
|
|
|
BMR = 655 + (4.35 × weight in lbs) + (4.7 × height in inches) – (4.7 × age in years) |
|
|
|
|
|
For men: |
|
|
|
|
|
BMR = 66 + (6.23 × weight in lbs) + (12.7 × height in inches) – (6.8 × age in years) |
|
|
|
|
|
2. **Harris-Benedict Equation:** This builds on BMR to estimate **Total Daily Energy Expenditure (TDEE)** by multiplying the BMR by an activity factor (e.g., 1.2 for sedentary, 1.55 for moderate activity). |
|
|
|
|
|
3. **Clinical Methods:** More accurate tests, like **indirect calorimetry**, measure oxygen consumption and carbon dioxide production to calculate actual energy expenditure. |
|
|
|
|
|
4. **Wearables and Apps:** Fitness trackers and smart devices offer real-time estimates, though they are less precise. |
|
|
|
|
|
Always consult a healthcare provider for personalized metabolic assessments or weight management planning. |
|
|
|
