| Disease | Task | Open-source Model | Accuracy Before Fine-tuning | Accuracy After Fine-tuning | Contact Support |
|---|---|---|---|---|---|
| Alzheimer disease | Predict Alzheimer disease | Llama3.1-8B | >90% | >90% | Contact |
| Frontotemporal dementia | Predict frontotemporal dementia | Llama3.1-8B | >90% | >90% | Contact |
| Lewy body dementia | Predict Lewy body dementia | Llama3.1-8B | >30% | >90% | Contact |
| Creutzfeldt-Jakob disease | Predict Creutzfeldt-Jakob disease | Llama3.1-8B | >50% | >90% | Contact |
| Mild cognitive impairment | Predict Mild cognitive impairment | Llama3.1-8B | >90% | >90% | Contact |
| Ischemic stroke | Predict Ischemic stroke | Llama3.1-8B | >80% | >90% | Contact |
| Parkinson disease | Predict Parkinson disease | Llama3.1-8B | >90% | >90% | Contact |
This table highlights some of the diseases for which we have fine-tuned the smaller Llama3.1-8B (8 billion parameters) model. The model's baseline prediction accuracy—prior to fine-tuning—ranges widely, from 20% to 100%. However, after fine-tuning with synthetic patient data, the Llama3.1-8B model consistently achieves >90% accuracy in predicting target diseases across various patient cases.
Due to the urgent need for clinical evidence supporting GenAI, you can swiftly evaluate one of these fine-tuned models in your real-world clinical settings (see GenAI Research Ideas) and potentially publish your findings in top journals like JAMA, NEJM, and Nature. Once validated with your real-world data, your fine-tuned model can be routinely used to assist in predictive healthcare.
If the table does not include the diseases or healthcare tasks that interest you, feel free to contact us, and we will create custom fine-tuned Llama3.1 models for you.
With our LLM technical support, any clinical team can deploy and fine-tune open-source LLMs to reap the benefits of GenAI. The Llama3.1-8B model requires only a single GPU server for both running and fine-tuning.
For well-funded clinical teams, the mid-size Llama3.1-70B model (70 billion parameters) is another excellent option for fine-tuning. Fine-tuned Llama3.1-70B models will be provided on demand, at the customer's request.