How do I train YOLO on my own custom dataset?

Question

Muhammad Rizwan Munawar · Accepted Answer

To train YOLO on a custom dataset: label your images in YOLO format, organise them into train/val folders, write a data.yaml describing the paths and class names, then call model.train() starting from a pretrained checkpoint. Validate and export when the metrics look right. 1. Label your images Draw boxes and assign classes using a tool like Label Studio, Roboflow, or CVAT, and export in YOLO format. Each image gets a .txt file with one line per object: class x_center y_center width height, all normalised 0-1. 2. Organise the folders datasets/mydata/
  images/train/  images/val/
  labels/train/  labels/val/ 3. Write data.yaml path: ../datasets/mydata
train: images/train
val: images/val
names:
  0: person
  1: forklift 4. Train from pretrained weights from ultralytics import YOLO

model = YOLO("yolo11n.pt")          # start from pretrained, not scratch
model.train(data="data.yaml", epochs=100, imgsz=640, batch=16)

metrics = model.val()              # check mAP before trusting it
model.export(format="onnx")        # deploy-ready format Arguments worth knowing imgsz: raise to 1280 for small objects; keep 640 for speed. epochs: 100 is a reasonable start; use early results to decide if more helps. batch: as large as your GPU memory allows; -1 lets Ultralytics auto-pick. model size: yolo11n/s for speed, yolo11m/l/x for accuracy. If accuracy disappoints after training, the fix is almost always in the data, not the training command.

How do I train YOLO on my own custom dataset?

1. Label your images

2. Organise the folders

3. Write data.yaml

4. Train from pretrained weights

Arguments worth knowing

Related questions

Hands-on tutorials

Stuck on this in a real project?