wellness-mini

Wellness-Mini: Mental Health Conversational AI

Model Description

Wellness-Mini is a fine-tuned version of SmolLM-135M, specifically adapted for mental health conversations and assessments.

Usage

In Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer
Load model and tokenizer
model = AutoModelForCausalLM.from_pretrained("ombhojane/wellness-mini", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("ombhojane/wellness-mini")
Example usage
messages = [{"role": "user", "content": "How are you feeling today?"}]
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(inputs, max_new_tokens=100)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

Via Pipeline

from transformers import pipeline
Create pipeline
pipe = pipeline(
"text-generation",
model="ombhojane/wellness-mini",
tokenizer="ombhojane/wellness-mini",
trust_remote_code=True
)
Generate text
response = pipe("How are you feeling today?", max_new_tokens=100)
print(response[0]['generated_text'])

Training Details

• Base model: SmolLM-135M
• Fine-tuned using supervised fine-tuning (SFT)
• Trained with both mental health assessment capabilities and proper identity responses

Limitations

• This is an AI assistant and not a replacement for professional medical advice
• Should be used as a supplementary tool only
• May not be suitable for emergency situations
• Responses should be verified by healthcare professionals

Intended Use

This model is intended to be used as a conversational AI assistant focused on mental health support and assessment. It can:

• Provide supportive responses to mental health concerns
• Help identify potential mental health indicators
• Engage in wellness-focused conversations

Bias and Risks

• May reflect biases present in training data
• Should not be used as sole diagnostic tool
• Responses should be reviewed by healthcare professionals

Creator

This model was developed by Sentinet AI Systems.