0

I am using YOLOV8 for instance segmentation using Google Colab. Now, after I download, train, validate, and test my dataset from Roboflow, I want to upload my model/weights to a specific project version to use it as my API. I follow their documentation upon deploying it, but I got an error due to the method is only for object detection type. I would really appreciate your attention.

This is my code looks like

and this is what it looks like when I open the link. message upon clicking the link

Code Snippet: project.version(2).deploy(model_type="yolov8", model_path=f"/content/drive/MyDrive/PATH/model/runs/segment/train4/")

Error Snippet: { "message": "Dataset is not of the correct type.", "type": "InvalidDatasetException", "hint": "This method can only be used on object detection datasets." }

I already applied the methods in the following articles but still I got the same error. https://roboflow.com/deploy https://docs.roboflow.com/upload-weights

Tim Cuizon
  • 3
  • 3

1 Answers1

1

Model weights uploads are only available for object detection models from YOLOv8 at this time. I've just made an update to the blog post to make it a bit more clear.

I do want to add that it was already added to the documentation when the feature was released.

Mohamed Traore
  • 191
  • 3
  • I see, thank you for this. What alternative do you reccomend for this? – Tim Cuizon Jan 26 '23 at 02:06
  • We’re working on adding it. If you want to get on the wait list & potentially early access when it’s ready please reach out (either via our forum or sales form depending on whether it’s for a personal or business project). – Brad Dwyer Jan 26 '23 at 13:31