Could you please share me a demo?
The repository is https://github.com/dmlc/gluon-cv . I am doing the semantic segmentation task. Could anybody please share me a demo?
There are two types of finetuning:
- Finetune from COCO trained segmentation model
- Finetune from ImageNet-trained backbone
The first one is preferred only if your own dataset has similar classes with COCO’s. In this case the link from @safrooze is the one you need: https://gluon-cv.mxnet.io/build/examples_segmentation/voc_sota.html
Otherwise it is better to finetune from a pretrained backbone directly, and the tutorials are: