Phi-3.5 Medical Q&A
Fine-tuned version of Phi-3.5-mini-instruct on synthetic medical Q&A data.
Training
- Base model: microsoft/Phi-3.5-mini-instruct
- Method: LoRA (r=16, alpha=32)
- Dataset: 126 synthetic medical Q&A pairs across 18 topics
- Epochs: 5 | Final loss: 0.20
- Hardware: Google Colab T4 GPU (free tier)
Topics covered
Chest pain, shortness of breath, diabetes management, thyroid panel, lab interpretation, cancer screening, heart health diet, and more.
- Downloads last month
- 6
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for gourikam/phi3-medical-qa
Base model
microsoft/Phi-3.5-mini-instruct