from sample.model import MyModel from visualizer import Visualizer import pickle import numpy as np import torch from imglib.image_io import imshow img_np = np.random.randint(0, 256, (100, 100, 3)) / 255 img = torch.tensor(img_np) model = MyModel() visualizer = Visualizer(model) visualizer.separate_layers() visualizer.describe_layers() output = visualizer.check_output(input("Please input layer name >>> "), img) output = np.array(output).astype(np.uint8) print(img_np.shape, output.shape) # imshow(img_np * 255, "hoge") # imshow(output * 255, "hoge")