Laughing

hi @onuralpszr, now the following workflow has been supported in ultralytics==8.0.5, please upgrade and have a try. :) ```python from ultralytics import YOLO model = YOLO("yolov8n.pt") # load a pretrained...

@onuralpszr closing the issue. Feel free to reopen it if you have any related issues. :)

@knoppmyth hi yeah actually we're using hydra but it seems not to support list this way. But we'll work on it. For now I recommend you to use python interface...

@knoppmyth hi now the following command is working as expect: ```bash yolo task=detect mode=export model=runs/detect/train/weights/last.pt imgsz=192,640 simplify=true format=onnx opset=11 ``` closing this issue, feel free to reopen it if you...

@AyushExel could be .. @AdGw can you try it again with `v5loader=True`?

@wvalcke hi can you use the latest version? I think `plot=True` should only work when it's the final epoch. [https://github.com/ultralytics/ultralytics/blob/d0b616e41e827e1ac65ee99da1339e707a984989/ultralytics/yolo/engine/validator.py#L90](https://github.com/ultralytics/ultralytics/blob/d0b616e41e827e1ac65ee99da1339e707a984989/ultralytics/yolo/engine/validator.py#L90) and yes it looks like a threading issue, we plot...

@VdLMV yeah the TAL assigner is not as efficient as yolov5 when handling the growth of instance number. Let's say there're 3 images which the number of labels of each...

@AyushExel @glenn-jocher actually I think we should also implement yolov5's label assignment, which is less accurate but more efficient. Then users are able to chose what works best for them.

@SyGoing hi, please provide more info, such as ultralytics package version and your training command. Also please show us the visualized batch data if you don't mind. :)

@SyGoing can't reproduce your issue with `coco128.yaml`, I just replaced your data file to coco128.yaml but my visualized image looks good. my script: ```python from ultralytics.yolo.v8 import detect def train_v8_detect():...