I have pruned a yolov8 model using : from torch.nn.utils import prune and the l1.structured method. I obtain a pruned_model.pt file. The pruned model's size was bigger than my original model. This might be because the models don't have the same format. The original model is a yolo model while the pruned one is a torch.save model. I would like to find a way to make sure that my pruned model has less weights and less layers than my original one.
I tried making two scripts that allow me to visualize the layers of each one of my models. The models not having the same architecture, this didn't allow me ton conclude on the effect of my pruning.