Configure and launch custom YOLOv11 training jobs.
Field | Purpose | Example |
---|---|---|
Name | Identifies the run in history | Hand_Detector_v1_10Ep |
Description | (optional) Notes for future you | Testing larger image size |
Variant | Params (M) | COCO mAP50-95² | Notes |
---|---|---|---|
n | 2.6 | 39.5 | Ultra-light, mobile & IoT |
s | 9.4 | 47.0 | Good for edge GPUs |
m | 20.1 | 51.5 | Balanced; default |
l | 26.2 | 54.0 | Higher accuracy, more VRAM |
x | 58.8 | 56.8 | Maximum accuracy, slowest |
m
is an excellent starting point for most users.Option | When to use |
---|---|
From Scratch | Fresh dataset or new architecture. |
Upload Weights (coming soon) | Resume / fine-tune a previous model. |
Parameter | UI Control | Range | What it does |
---|---|---|---|
Epochs | Slider | 1-100 | Number of full passes through the training set. |
Image Size | Radio-buttons (320 / 640) | 320 or 640 px | Resolution the model is trained with. Higher helps small objects but needs more VRAM. |