I was using an anomaly detection framework called GLASS ( https://github.com/cqylunlun/GLASS ). After I've trained on my own dataset, GLASS returns the weights of the best epoch on a .pth file. At this point, I'd like to perform the inference on the trained model, but before I'd load the trained model and I assume using the .pth file, but I was reading I also need to build again the GLASS class which is also based on a backbone like resnet. Can any help me further? Continue reading...
