SCI Assistant 7B

A specialized language model for spinal cord injury (SCI) information and support, based on OpenHermes-2.5-Mistral-7B with custom LoRA fine-tuning.

Model Description

This model has been fine-tuned specifically to provide accurate, helpful information about spinal cord injuries, including:

Medical information about SCI conditions and symptoms
Practical advice for daily living with SCI
Equipment recommendations for wheelchairs, adaptive technology, etc.
Exercise and rehabilitation guidance
Emotional support and community resources

Training Data

The model was trained on curated SCI-related content including:

Medical literature and research papers
Patient education materials
Community forums and discussions
Rehabilitation guides and resources

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("your-username/sci-assistant-7b")
tokenizer = AutoTokenizer.from_pretrained("your-username/sci-assistant-7b")

# Example usage
prompt = "What are the signs of autonomic dysreflexia?"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_length=200)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)

Intended Use

Educational purposes - Learning about SCI conditions and management
Community support - Providing accessible information to SCI community
Research - Supporting SCI-related research and development

Limitations

This model provides educational information only
Always consult healthcare professionals for medical advice
Not a replacement for professional medical care
May not reflect the most recent medical developments

Technical Details

Base Model: teknium/OpenHermes-2.5-Mistral-7B
Fine-tuning: LoRA (Low-Rank Adaptation)
Parameters: ~7 billion
Precision: FP16

License

Please respect the original OpenHermes-2.5 license terms.

Acknowledgments

Built on the excellent OpenHermes-2.5-Mistral-7B model by Teknium. Training data curated from publicly available SCI educational resources.