NeuroJenML SFT LoRA Adapter
This is the LoRA adapter fine-tuned on MedGemma-4B-it. Designed to predict central pathology from peripheral Alzheimer's findings.
Training Details
- Base Model:
google/medgemma-4b-it - Training Loss:
0.0878 - Training Epochs:
3 - Training Examples:
1747 - Validation Loss:
0.059457048773765564 - Adapter Configuration: LoRA (r=16, alpha=16)
- Downloads last month
- 887
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support