J

thanks for sharing

Well, I think it works with `model.load_state_dict` ``` # weights = torchvision.models.EfficientNet_B0_Weights.DEFAULT efficientnet_b0_path = 'models/efficientnet_b0_rwightman-3dd342df.pth' model = torchvision.models.efficientnet_b0() for param in model.parameters(): param.required_grad = False model.load_state_dict(torch.load(efficientnet_b0_path)) ``` ``` summary(model=model, input_size=(32,...