metadata
language: en
license: apache-2.0
library_name: transformers
tags:
- chatbot
- caregiver-support
- autism-support
- conversational
- mental-health
- support-chatbot
pipeline_tag: text-generation
widget:
- text: |
### Question:
I'm feeling overwhelmed with my autistic child's needs
### Answer:
example_title: Caregiver Support
- text: |
### Question:
My child is having a meltdown, what should I do?
### Answer:
example_title: Meltdown Support
- text: |
### Question:
I'm exhausted and don't know what to do
### Answer:
example_title: Caregiver Burnout
Rafiki Chatbot 🤗
A supportive AI assistant for caregivers of autistic children
🌟 Overview
Rafiki Chatbot is a fine-tuned language model specifically designed to provide emotional support, practical advice, and compassionate responses to caregivers of autistic children. The name "Rafiki" means "friend" in Swahili, reflecting the chatbot's supportive and friendly nature.
🎯 Purpose
This model helps caregivers who are:
- Feeling overwhelmed or exhausted
- Seeking strategies for managing challenging behaviors
- Looking for emotional support and validation
- Needing practical advice for daily caregiving challenges
- Seeking understanding and community
🏗️ Model Details
- Base Model: microsoft/phi-3-mini-4k-instruct
- Fine-tuning Method: LoRA (Low-Rank Adaptation)
- Training Data: Custom caregiver support conversations
- Language: English
- Context Length: 4,000 tokens
📊 Training Data
The model was fine-tuned on a carefully curated dataset including:
- Caregiver support conversations
- Autism parenting strategies
- Emotional support dialogues
- Crisis management scenarios
- Daily routine optimization tips
🚀 Usage
Using the Inference API
from transformers import pipeline
# Initialize the chatbot
chatbot = pipeline(
"text-generation",
model="Tizzu668/rafiki-chatbot",
device_map="auto"
)
# Example conversation
prompt = "### Question:\nI'm feeling overwhelmed with my autistic child's constant needs\n### Answer:\n"
response = chatbot(
prompt,
max_new_tokens=128,
temperature=0.7,
do_sample=True
)
print(response[0]['generated_text'])