# Training Resolutions by Model Type

학습 해상도는 모델 정확도, 추론 속도 및 학습 시간에 영향을 줍니다. 각 모델 아키텍처는 이러한 요소들을 균형 있게 맞추는 기본 해상도를 가지고 있습니다. 기본적으로 Roboflow는 선택한 모델 아키텍처에 대한 기본 학습 해상도를 제안합니다. 

아래 표는 각 모델 아키텍처 및 크기에 대한 기본 학습 해상도를 보여줍니다. 새 Dataset Version을 만들 때 resize 전처리 단계를 구성하여 이러한 기본값을 재정의할 수 있습니다. [Dataset Version](https://docs.roboflow.com/roboflow/roboflow-ko/datasets/dataset-versions).

### Object Detection

<table><thead><tr><th>Model Type</th><th width="273.3359375">Family &#x26; Size</th><th>Default Training Resolution</th></tr></thead><tbody><tr><td>Object Detection</td><td>RF-DETR Nano</td><td>384×384</td></tr><tr><td>Object Detection</td><td>RF-DETR Small</td><td>512×512</td></tr><tr><td>Object Detection</td><td>RF-DETR Medium</td><td>576×576</td></tr><tr><td>Object Detection</td><td>RF-DETR Large</td><td>704×704</td></tr><tr><td>Object Detection</td><td>RF-DETR X Large</td><td>700x700</td></tr><tr><td>Object Detection</td><td>RF-DETR 2X Large</td><td>880x880</td></tr><tr><td>Object Detection</td><td>Roboflow 3.0 - Fast</td><td>640×640</td></tr><tr><td>Object Detection</td><td>Roboflow 3.0 - Accurate</td><td>640×640</td></tr><tr><td>Object Detection</td><td>Roboflow 3.0 - Medium</td><td>640×640</td></tr><tr><td>Object Detection</td><td>Roboflow 3.0 - Large</td><td>640×640</td></tr><tr><td>Object Detection</td><td>Roboflow 3.0 - Extra Large</td><td>640×640</td></tr><tr><td>Object Detection</td><td>YOLOv26(n/s/m/l/x)</td><td>640×640</td></tr><tr><td>Object Detection</td><td>YOLOv12 (n/s/m/l/x)</td><td>640×640</td></tr><tr><td>Object Detection</td><td>YOLOv11 (n/s/m/l/x)</td><td>640×640</td></tr><tr><td>Object Detection</td><td>YOLOv10 (n/s/m/b/l/x)</td><td>640×640</td></tr><tr><td>Object Detection</td><td>YOLOv9 (s/m/c/e)</td><td>640×640</td></tr><tr><td>Object Detection</td><td>YOLOv8 (n/s/m/l/x)</td><td>640×640</td></tr><tr><td>Object Detection</td><td>YOLOv5 (n/s/m/l/x)</td><td>640×640</td></tr><tr><td>Object Detection</td><td>YOLOv7 (legacy)</td><td>640×640</td></tr><tr><td>Object Detection</td><td>YOLO‑NAS Small</td><td>640×640</td></tr><tr><td>Object Detection</td><td>YOLO‑NAS Medium</td><td>640×640</td></tr><tr><td>Object Detection</td><td>Roboflow Instant</td><td>1008x1008</td></tr></tbody></table>

### Instance Segmentation

<table><thead><tr><th>Model Type</th><th width="272.8203125">Family &#x26; Size</th><th>Default Training Resolution</th></tr></thead><tbody><tr><td>Instance Segmentation</td><td>RF-DETR Seg Nano</td><td>312x312</td></tr><tr><td>Instance Segmentation</td><td>RF-DETR Seg Small</td><td>384x384</td></tr><tr><td>Instance Segmentation</td><td>RF-DETR Seg Medium</td><td>432x432</td></tr><tr><td>Instance Segmentation</td><td>RF-DETR Seg Large</td><td>504x504</td></tr><tr><td>Instance Segmentation</td><td>RF-DETR Seg X Large</td><td>624x624</td></tr><tr><td>Instance Segmentation</td><td>RF-DETR Seg 2X Large</td><td>768x768</td></tr><tr><td>Instance Segmentation</td><td>Roboflow 3.0 - Fast (Seg)</td><td>640×640</td></tr><tr><td>Instance Segmentation</td><td>Roboflow 3.0 - Accurate (Seg)</td><td>640×640</td></tr><tr><td>Instance Segmentation</td><td>Roboflow 3.0 - Medium (Seg)</td><td>640×640</td></tr><tr><td>Instance Segmentation</td><td>Roboflow 3.0 - Large (Seg)</td><td>640×640</td></tr><tr><td>Instance Segmentation</td><td>Roboflow 3.0 - Extra Large (Seg)</td><td>640×640</td></tr><tr><td>Instance Segmentation</td><td>YOLO-seg (v8/10/11/12)</td><td>640×640</td></tr><tr><td>Instance Segmentation</td><td>SAM3 (Segment Anything 3)</td><td>1008x1008</td></tr><tr><td>Instance Segmentation</td><td>Semantic segmentation (DeepLabV3+)</td><td>≥ 512×512</td></tr></tbody></table>

### Classification & Pose

<table><thead><tr><th>Model Type</th><th width="272.66796875">Family &#x26; Size</th><th>Default Training Resolution</th></tr></thead><tbody><tr><td>Classification &#x26; Pose</td><td>Resnet-18/34/50</td><td>224x224</td></tr><tr><td>Classification &#x26; Pose</td><td>YOLO-cls (v8/11)</td><td>224x224</td></tr><tr><td>Classification &#x26; Pose</td><td>Vision Transformer (ViT)</td><td>224x224</td></tr><tr><td>Classification &#x26; Pose</td><td>YOLO-pose (keypoints)</td><td>640x640</td></tr></tbody></table>

### Multimodal/VLM

<table><thead><tr><th>Model Type</th><th width="272.96484375">Family &#x26; Size</th><th>Default Training Resolution</th></tr></thead><tbody><tr><td>Multimodal/VLM</td><td>PaliGemma 2 - 3 B</td><td>448x448</td></tr><tr><td>Multimodal/VLM</td><td>PaliGemma 2 - 10 B/28 B</td><td>448x448</td></tr><tr><td>Multimodal/VLM</td><td>Florence-2</td><td>448x448</td></tr><tr><td>Multimodal/VLM</td><td>QWEN 2.5 VL</td><td>448x448</td></tr><tr><td>Multimodal/VLM</td><td>SmolVLM2</td><td>384x384</td></tr></tbody></table>