training and validation

[bbetula]

The training code of this model seems to call YOLO for training. Shouldn't the trained model be a YOLOv8n-seg type model, rather than a FastSAM type model?

see this
from ultralytics import YOLO
model = YOLO(model="yolov8x-seg.yaml",
)
model.train(data="sa.yaml",
epochs=100,
batch=32,
imgsz=1024,
overlap_mask=False,
save=True,
save_period=5,
device='0,1,2,3,4,5,6,7',
project='fastsam',
name='test',
val=False,)

[harshnandwana]

no that is correct as it requires yaml file where as your understanding is because of the folder structure having pt