# %% [code] # %% [markdown] # # Prepare data and environement # %% [code] {"execution":{"iopub.status.busy":"2021-12-20T11:18:32.226832Z","iopub.execute_input":"2021-12-20T11:18:32.227218Z","iopub.status.idle":"2021-12-20T11:19:10.914635Z","shell.execute_reply.started":"2021-12-20T11:18:32.227187Z","shell.execute_reply":"2021-12-20T11:19:10.913263Z"}} import os import zipfile test_folder="./input/test1" train_folder= "./input/train" # Extract data if not os.path.isdir(test_folder): with zipfile.ZipFile("../input/dogs-vs-cats/test1.zip","r") as zip_ref: zip_ref.extractall("./input/") if not os.path.isdir(train_folder): with zipfile.ZipFile("../input/dogs-vs-cats/train.zip","r") as zip_ref: zip_ref.extractall("./input/") output = os.popen('pip install keras-visualizer').read() output += os.popen('pip install graphviz').read() print(output) from graphviz import Source from keras_visualizer import visualizer # Put cats and dogs in repository import glob, os, shutil for item_class in["cat","dog"]: files = glob.glob(f'{train_folder}/{item_class}.*.jpg') class_folder=os.path.join(train_folder,item_class) os.makedirs(class_folder, exist_ok=True) for file in files: shutil.move(file, class_folder) # %% [code] {"execution":{"iopub.status.busy":"2021-12-20T11:19:18.727824Z","iopub.execute_input":"2021-12-20T11:19:18.728161Z","iopub.status.idle":"2021-12-20T11:19:22.851989Z","shell.execute_reply.started":"2021-12-20T11:19:18.728123Z","shell.execute_reply":"2021-12-20T11:19:22.850518Z"}} # Utililty function def list_files(startpath): from PIL import Image import os for root, dirs, files in os.walk(startpath): level = root.replace(startpath, '').count(os.sep) indent = ' ' * 4 * (level) print('{}{}/'.format(indent, os.path.basename(root))) subindent = ' ' * 4 * (level + 1) for itf, f in enumerate(files): if itf <6: if f.split(".")[-1] == "jpg": image = Image.open(os.path.join(root,f)) f = f"{f.ljust(30,' ')} size: {image.size}" print('{}{}'.format(subindent, f)) else: print(subindent,"6 files shown.. on",len(files),"\t", len(files)-6, "files left") break def save_model_as_file(model,filename,format="png"): dotfile = f"{filename}.dot" visualizer(model, filename=dotfile, format=None, view=False) print(f"Model exported as,", dotfile) s = Source.from_file(dotfile) s.format=format s.render(filename) print(f"Model exported as {format} saved as {filename}.{format}") # Plot an array of things # [{"title":"filename"}] # [{"title":"2darray"}] # [{"title":"pilimg"}] def plot_figures(images_dicts, ncols=5): import matplotlib.pyplot as plt import numpy as np import PIL from PIL import Image # settings import math nrows = math.ceil(len(images_dicts) / ncols) figsize = [nrows*6, ncols * 1.5] # figure size, inches # prep (x,y) for extra plotting on selected sub-plots xs = np.linspace(0, 2 * np.pi, 60) # from 0 to 2pi ys = np.abs(np.sin(xs)) # absolute of sine # create figure (fig), and array of axes (ax) fig, ax = plt.subplots(nrows=nrows, ncols=ncols, figsize=figsize) if len(images_dicts) >= ncols: ax=ax.flatten() # plot simple raster image on each sub-plot for axi, img_dict in zip(ax, images_dicts): title, img = list(img_dict.items())[0] print("Plotting",title) if type(img) is str: img = Image.open(img) if type(PIL.PngImagePlugin.PngImageFile) is PIL.PngImagePlugin.PngImageFile: img = np.array(img) # i runs from 0 to (nrows*ncols-1) # axi is equivalent with ax[rowid][colid] axi.imshow(img) axi.set_title(title) # one can access the axes by ax[row_id][col_id] # do additional plotting on ax[row_id][col_id] of your choice # ax[0][2].plot(xs, 3 * ys, color='red', linewidth=3) # ax[4][3].plot(ys ** 2, xs, color='green', linewidth=3) plt.tight_layout() plt.show() files=glob.glob(f'{train_folder}/cat/*.jpg')+glob.glob(f'{train_folder}/dog/*.jpg') import random import os.path random.shuffle(files) files=files[:30] some_figs=[ {os.path.basename(file):file} for file in files ] plot_figures(some_figs) # %% [markdown] # # One block code from tutorial [github](https://gist.github.com/fchollet/0830affa1f7f19fd47b06d4cf89ed44d) # %% [code] from keras.preprocessing.image import ImageDataGenerator from keras.models import Sequential from keras.layers import Conv2D, MaxPooling2D from keras.layers import Activation, Dropout, Flatten, Dense from keras import backend as K # dimensions of our images. img_width, img_height = 150, 150 train_data_dir = train_folder validation_data_dir = train_folder nb_train_samples = 2000 nb_validation_samples = 800 epochs = 50 batch_size = 16 if K.image_data_format() == 'channels_first': input_shape = (3, img_width, img_height) else: input_shape = (img_width, img_height, 3) model = Sequential() model.add(Conv2D(32, (3, 3), input_shape=input_shape)) model.add(Activation('relu')) model.add(MaxPooling2D(pool_size=(2, 2))) model.add(Conv2D(32, (3, 3))) model.add(Activation('relu')) model.add(MaxPooling2D(pool_size=(2, 2))) model.add(Conv2D(64, (3, 3))) model.add(Activation('relu')) model.add(MaxPooling2D(pool_size=(2, 2))) model.add(Flatten()) model.add(Dense(64)) model.add(Activation('relu')) model.add(Dropout(0.5)) model.add(Dense(1)) model.add(Activation('sigmoid')) model.compile(loss='binary_crossentropy', optimizer='rmsprop', metrics=['accuracy']) save_model_as_file(model,"my_mod",format="png") # this is the augmentation configuration we will use for training train_datagen = ImageDataGenerator( rescale=1. / 255, shear_range=0.2, zoom_range=0.2, validation_split=0.5, horizontal_flip=True) # this is the augmentation configuration we will use for testing: # only rescaling test_datagen = ImageDataGenerator(rescale=1. / 255, validation_split=0.5 ) train_generator = train_datagen.flow_from_directory( train_data_dir, target_size=(img_width, img_height), batch_size=batch_size, class_mode='binary', subset="training", ) validation_generator = test_datagen.flow_from_directory( validation_data_dir, target_size=(img_width, img_height), batch_size=batch_size, class_mode='binary', subset="validation", ) model.fit( train_generator, steps_per_epoch=nb_train_samples // batch_size, epochs=epochs, validation_data=validation_generator, validation_steps=nb_validation_samples // batch_size) model.save_weights('first_try.h5') # %% [code]