Problem
Users need a guided workspace for LoRA fine-tuning instead of editing raw configuration.
Scope
- Add LoRA controls for rank, alpha, dropout, target modules, epochs, batch size, and learning rate.
- Use sliders and bounded inputs for numeric settings.
- Show estimated training impact as settings change.
Acceptance Criteria
- Users can configure a LoRA run from the TuneOS UI.
- Settings are validated before training starts.
- Configuration can be saved to the active project.
Problem
Users need a guided workspace for LoRA fine-tuning instead of editing raw configuration.
Scope
Acceptance Criteria