def main(): ######################################################################## # Loads the configuration for the experiment from the configuration file #config, learning_rate, batch_size, num_epochs, target_classes = load_config('configuration.json') ######################################################################## # Obtain the PyTorch data loader objects to load batches of the datasets #train_loader, val_loader, test_loader, classes = get_data_loader(target_classes, batch_size) # Load the model net = torch.load("./saved_models/final.pt") val_images = np.load("./data/test_images_2.npy") val_label = np.load("./data/test_label_2.npy") val_dataset = MyDataset(val_images, val_label) val_loader = DataLoader(val_dataset, batch_size=1024, shuffle=False) #model_path = get_model_name(config) #net.load_state_dict(torch.load(model_path)) # Evaluate the model for the confusion matrix #evaluate_confusion_matrix(model_path, net, test_loader, target_classes) # Visualize examples for false positive / true negative #evaluate_visual_confusion_matrix('Plot6', net, val_loader, ['no landing pad','landing pad']) #cm=evaluate_confusion_matrix('Plot4', net, val_loader, ['no landing pad','landing pad']) #plot_confusion_matrix('Plot5', cm, ['no landing pad','landing pad'],normalize=False,title='Confusion matrix',cmap=plt.cm.Blues) print(evaluate(net, val_loader))
def load_data(batch_size): train_path = "data/rand_flowers_40/train/" train_path_masked = "data/rand_flowers_40_masked/train/" val_path = "data/rand_flowers_40/val/" val_path_masked = "data/rand_flowers_40_masked/val/" train_dataset = MyDataset(train_path, train_path_masked) val_dataset = MyDataset(val_path, val_path_masked) train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True) val_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=False) return train_loader, val_loader
def load_data(batch_size): train_images = np.load("./data/train_images.npy") val_images = np.load("./data/val_images.npy") #test_images=np.load("./data/test_images.npy") train_label = np.load("./data/train_label.npy") val_label = np.load("./data/val_label.npy") #test_label =np.load("./data/test_label.npy") train_dataset = MyDataset(train_images, train_label) val_dataset = MyDataset(val_images, val_label) #test_dataset = MyDataset(test_images, test_label) train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True) val_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=False) #test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False) return train_loader, val_loader #, test_loader
def load_data(batch_size): test_path = "data/rand_flowers_40/test/" test_path_masked = "data/rand_flowers_40_masked/test/" test_path = "data/brick_40/test/" test_path_masked = "data/brick_40_masked/test/" test_dataset = MyDataset(test_path, test_path_masked) test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False) return test_loader
c = np.transpose(c, (1, 0, 2)) plt.imshow(c) plt.gray() plt.show() # print(img.size) # (4032, 2268) model = torch.load("./data/model2.pt") feat_valid = np.load("./data/images_test.npy") label_valid = np.load("./data/labels_test.npy") stride = 252 # feat_valid = feat_valid.transpose((0,2,1)) batch_size = 1 valid_dataset = MyDataset(X=feat_valid, y=label_valid) val_loader = DataLoader(valid_dataset, batch_size=batch_size, shuffle=False) total_corr = 0 matrix = np.zeros(img.size) for i, data in enumerate(val_loader): images, label = data images = images.float() # print(label) # print("data: ",inputs.shape) # print("label",label.detach().numpy()[0][0]) # print("label",label.detach().numpy()[0][1]) outputs = model(images) # print("outputs:{}".format(outputs.item()))
# cv2.destroyAllWindows() # cv2.imwrite(newdir + "/" + filename, frame) print("size:",new_img.shape) crop_image.append(new_img) index.append((i, j)) j += stride i += stride images_zeros = np.array(crop_image) label_zeros = np.array(index) batch_size = 1 valid_dataset = MyDataset(X=images_zeros, y=label_zeros) val_loader = DataLoader(valid_dataset, batch_size=batch_size, shuffle=False) total_corr = 0 matrix_0 = np.zeros((newframe.shape[0] // stride , newframe.shape[1] // stride)) ### 这里 # matrix_0 = np.zeros((newframe.shape[0], newframe.shape[1] )) ### 这里 matrix = np.zeros((newframe.shape[0] // stride , newframe.shape[1] // stride )) # matrix = np.zeros((newframe.shape[0], newframe.shape[1])) for i, data in enumerate(val_loader): images, label = data print("label", label) images = images.float() print(images.shape) outputs = model(images)