Laughing

@AyushExel @glenn-jocher adding TODO flag

@guishilike hi this issue has solved in this PR #699 which probably will be merged later today. :)

@lzt9746 hi thanks for the reporting. I think I know why but actually we haven't finished our x6 model training, so I got no x6 model to test for now....

@lzt9746 ok thanks! so it it model.stride issue. we'll fix it later. :) > Besides, you can run these code to get your released yolov8x6 model, code will download model...

@lzt9746 I've made a PR to fix it. #407

@AyushExel I think we haven't added this arg right?

@lonnylundsten hi the `classes` has been implemented which you can use it by: ```bash yolo segment predict model=yolov8n-seg.pt classes=0,1 ``` closing the issue, feel free to reopen it if you...

@RelativelyFine By default it'll save to `runs/task/predict` and the `task` can be `detect/segment/classify`(depends on what task you're using). The save path consists of two args -> `project`/`name`, and you can...

@JiaxinWang123 ```python model = YOLO("yolov8n.pt") model.predict(source=..., project="xx", name="xxx") ``` then it'll be save in `xx/xxx`.

@CesareDavidePace yeah we have objectness branch in v5 so it doesn't need cls loss when there's only one class. But in v8 we removed the objectness branch so we actually...