Built-in training runtimes
The YOLO · RF-DETR · Pointcept (PTv3) runtimes the platform builds itself, and the data each one takes
Even without bringing in an external image, the platform has three training runtimes it builds and uses itself. They also run under the same contract as this section — built-in trainers follow no different rules. They make good reference implementations when you build a new trainer.
One framework = one runtime image. Frameworks with different dependencies are not packed into one image. You choose the framework when you create a variant, and cannot change it later.
| Framework | Task | Model variants (wizard step 3) | Data | GPU |
|---|---|---|---|---|
| YOLO (ultralytics) | Object detection | yolo26n · yolo12n · yolo12s · yolo11n · yolo11s | Images + annotations | Optional |
| Instance segmentation | yolo11n-seg · yolo11s-seg · yolo11m-seg | Images + polygon annotations | Optional | |
| RF-DETR | Object detection | rf-detr-nano · -small · -medium · -large | Images + annotations | Optional |
| Instance segmentation | rf-detr-seg-nano · -small · -medium | Images + polygon annotations | Optional | |
| Pointcept (PTv3) | Semantic segmentation | pt-v3 | PLY point clouds | Required |
What differs
| YOLO | RF-DETR | Pointcept | |
|---|---|---|---|
| Dockerfile | Dockerfile | Dockerfile.rfdetr | Dockerfile.pointcept (two-stage build) |
| Pinned versions | ultralytics 8.4.68 | rfdetr 1.9.4 | A specific Pointcept commit, torch 2.9.1 |
| Training data it builds | YOLO labels (images/labels/{train,val} + data.yaml) | COCO ({train,valid}/_annotations.coco.json, category ids from 1) | {train,val}/{scene}/{coord,color,segment}.npy |
| Label source | COCO · LabelMe · VOC annotations | Same | Integer property inside the PLY |
| Pretrained weight location | Working directory (/geo) | RF_HOME (/geo/work/weights) | — |
- All three trainers build a dataset in their own format in
GEO_WORK_DIR, and link images with symbolic links instead of copying them (/geo/datasetis read-only, so there is no reason to copy). Only point clouds are converted from.plyto.npy. - Labels without polygons are downgraded to bbox rectangles for segmentation training.
- Pointcept uses CUDA-only extensions, so submitting without a GPU is rejected at submission.
- The label property of a point cloud is the integer property whose name contains
label·class·segment·semantic, or, if there is none, the first integer property that is not a coordinate. A float property whose values are all integers is also read as an integer.
To use them
A built-in runtime is used for training only after its image is published (built inside the cluster and pushed to the registry). A tenant administrator creates and publishes a variant per framework in Training runtime (학습 런타임), or a system administrator hands one out as a shared image. The detailed procedure is in the Administration guide.