Studio Features
Key capabilities of AI Training Studio.
Text Generation
| Feature | Description |
|---|---|
| Experiment tab | Fine-tuning experiment list with New Experiment button |
| Filters | Name search, Status and Method dropdowns |
| Table | Status, Name, Methods, Steps, Owner, Duration, Created/Completed at, Actions |
| New Experiment | Queue Select → Training Settings → Hyperparameters form |
| Methods | SFT / DPO / CPT / GKD / GRPO |
| Training Mode | Full Finetuning / LoRA / QLoRA |
| Teacher | Activated when GKD is selected |
| Hyperparameters | Basic / Resource / Advanced / Expert sections |
| Go ML Flow | Link to ML Flow metrics |
| Cancel | Creating / Pending states only |
| Delete | Delete modal → completion toast |
| Pagination | 10 rows per page |
Tabular
| Feature | Description |
|---|---|
| Experiment list | Status, Name, Owner, Created at, Actions |
| New Experiment | Queue Select + dataset/parameter form |
| Detail page | Header card (with Local Queue), training metric results |
| Register Model drawer | Category (dropdown / Custom), Description → Model Registry |
Pipeline Builder
| Feature | Description |
|---|---|
| Local Queue selection | Top-tab dropdown, required |
| Canvas | Drag-and-drop node composition |
| Left panel | Storage / Run node types |
| Right panel | Instance Name, Disclosure groups, TextField, SelectBox |
| Empty state | CTA message |
| Submit | Pipeline submission button |