How to keep the class weight for coco 2017 when pre-train model

[bzha5848]

Search before asking

• [X] I have searched the YOLOv5 issues and found no similar feature requests.

Description

Hi there,

I'm currently doing fine-tuning with coco2017 as you used for weight, just wondering if there's any method or code to keep the coco class as the same weight while tran the novel class?

Thanks so much!!

Use case

No response

Additional

No response

Are you willing to submit a PR?

• [ ] Yes I'd like to help by submitting a PR!

[MartinPedersenpp]

The easiest way would be to add the coco2017 dataset to your novel dataset and train the model. This way you keep all the 80 coco classes and add your novel classes. I am pretty sure that no matter what you do, the weights of the layers will get skewed away from the classes in coco2017, if the images and annotations are not in your training data.

[glenn-jocher]

@bzha5848 see https://community.ultralytics.com/t/how-to-combine-weights-to-detect-from-multiple-datasets/38/55

[github-actions[bot]]

👋 Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 🚀 resources:

• Wiki – https://github.com/ultralytics/yolov5/wiki
• Tutorials – https://docs.ultralytics.com/yolov5
• Docs – https://docs.ultralytics.com

Access additional Ultralytics ⚡ resources:

• Ultralytics HUB – https://ultralytics.com/hub
• Vision API – https://ultralytics.com/yolov5
• About Us – https://ultralytics.com/about
• Join Our Team – https://ultralytics.com/work
• Contact Us – https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 🚀 and Vision AI ⭐!