Peterande
Instance Segmentation
Hi!
I was able to train and fine tune on custom datasets and the performance is far better than other SOTA models (lmk if anyone needs help, more than happy to assist). I am wondering whether there will be a release for the instance segmentation part or not. It would be helpful if there is clear instructions to update the architecture.
Hi,
Can the dataset format to train yolov11 be used to train d-fine?
Hi,
Can the dataset format to train yolov11 be used to train d-fine?
Yes, but you will have to make the category_id 0-indexed. Make sure you put num_classes as n+1 in the yaml file, where n is the number of classes in the dataset. This is assuming that your initial dataset is in coco format which yolo requires.
Hello, how did you fine-tune the model? The fine-tuning parameters need to be modified. My attempt to fine-tune the effect is not very ideal. Please help me very much thank you for your help.
Hello, how did you fine-tune the model? The fine-tuning parameters need to be modified. My attempt to fine-tune the effect is not very ideal. Please help me very much thank you for your help.
Please provide more information (e.g., errors you are getting, dataset etc.) so I can help you
Hi!
I was able to train and fine tune on custom datasets and the performance is far better than other SOTA models (lmk if anyone needs help, more than happy to assist). I am wondering whether there will be a release for the instance segmentation part or not. It would be helpful if there is clear instructions to update the architecture.
https://github.com/talebolano/Contourformer
Would really love it if you'd integrate the work from ContourFormer !!
Would make me pick d-fine over rf-detr
