Irutingabo/pregnancy-assistant-tinyllama

Pregnancy Healthcare Assistant

This model is a fine-tuned version of TinyLlama/TinyLlama-1.1B-Chat-v1.0 using LoRA (Low-Rank Adaptation) for pregnancy and maternal healthcare questions.

Model Description

• Base Model: TinyLlama-1.1B-Chat-v1.0
• Fine-tuning Method: LoRA (Parameter-Efficient Fine-Tuning)
• Training Data: 2,806 pregnancy and maternal health Q&A pairs
• Domain: Pregnancy, maternal health, prenatal care, labor, delivery, breastfeeding, postpartum

Intended Use

This model is designed to provide informational responses to pregnancy and maternal health questions. Important: This model should NOT replace professional medical advice. Always consult qualified healthcare providers for medical concerns.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import torch

# Load base model
base_model = AutoModelForCausalLM.from_pretrained("TinyLlama/TinyLlama-1.1B-Chat-v1.0")
tokenizer = AutoTokenizer.from_pretrained("TinyLlama/TinyLlama-1.1B-Chat-v1.0")

# Load fine-tuned adapter
model = PeftModel.from_pretrained(base_model, "Irutingabo/pregnancy-assistant-tinyllama")

# Generate response
prompt = "What are the signs of early labor?"
inputs = tokenizer(prompt, return_tensors="pt")
with torch.no_grad():
    outputs = model.generate(**inputs, max_new_tokens=200, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)

Limitations

• Specialized for pregnancy/maternal health domain only
• General knowledge outside this domain is limited
• Should not be used for medical diagnosis or treatment decisions
• Always recommend consulting healthcare professionals

Training Details

• Training method: LoRA fine-tuning
• Trainable parameters: 4.5M (0.41% of total model parameters)
• Training dataset: Custom pregnancy Q&A dataset
• Training framework: Transformers + PEFT