Train & Fine-Tune Models
Medical Reasoning via LoRA Fine-Tuning
Gemma-3-12B adapted for symptom triage, drug interactions, and lab interpretation: 7.8GB GGUF for fully offline clinical use.
The 4-step NEO workflow
- 1
Describe the task
Tell NEO the model family, adaptation goal, and target behaviour.
- 2
Add context for NEO
Share datasets, checkpoints, hardware budget, and quality constraints.
- 3
NEO implements & delivers
NEO runs training, evals, and hands back versioned artifacts and reports.
- 4
Follow up or test it out
Compare metrics, sweep hyperparameters, and promote the winning run.
Ask NEO
How to run this scenario
Run "Medical Reasoning via LoRA Fine-Tuning" in your workspace: NEO plans data prep, training runs, and evals so you get reproducible weights and configs, not one-off notebooks.
Approach
What NEO focuses on
- Ground the plan in your repo: datasets, targets, and hardware constraints
- Iterate on training with logged metrics, checkpoints, and clear artifacts
- Gate promotion with evals you can re-run on new data or model versions
Outcomes
What you get
- Versioned configs and checkpoints tied to experiments
- Reports you can share for quality, cost, and regression tracking
- A straight path from experiment to staging with inspectable outputs
Ready to try for yourself?
Open NEO in VS Code or Cursor and describe this scenario. NEO plans the work, runs experiments, and ships artifacts you can review and iterate on.