Tensorflow - Using the Model

Dec 29 2019

This is all taken from sentdex's excellent Tensorflow tutorial series. Check it all out here:
https://www.youtube.com/playlist?list=PLQVvvaa0QuDfhTox0AjmQ6tvTgMBZBEXN\

Now we can set up a file that uses our created model to predict an output when given some input.
I ended up going with a model that uses 3 Conv2D layers with 96 nodes per layer. I also changed the amount of epochs to 7 after some trial and error, this seemed like an ok trade off between acc and val_acc before things started to overfit. The val_acc and val_loss aren't great but not terrible after the amount of tests I ran.
The model is found below:

import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout, Activation, Flatten, Conv2D, MaxPooling2D
from tensorflow.keras.callbacks import TensorBoard
import pickle
import time
import numpy as np


x = np.asarray(pickle.load(open("x.pickle", "rb")))
y = np.asarray(pickle.load(open("y.pickle", "rb")))

x = x/255.0

dense_layers = [0]
layer_sizes = [96]
conv_layers = [3]
dense_layer_size = 512

for dense_layer in dense_layers:
    for layer_size in layer_sizes:
        for conv_layer in conv_layers:
            NAME = f"96-layer_size-{int(time.time())}"
            tensorboard = TensorBoard(log_dir=f"logs\\{NAME}")

            model = Sequential()
            model.add(Conv2D(layer_size, (3,3), input_shape=x.shape[1:]))
            model.add(Activation("relu"))
            model.add(MaxPooling2D(pool_size=(2,2)))

            for l in range(conv_layer - 1):
                model.add(Conv2D(layer_size, (3,3)))
                model.add(Activation("relu"))
                model.add(MaxPooling2D(pool_size=(2,2)))

            model.add(Flatten())  # This converts our 3D feature maps to 1D feature vectors

            for l in range(dense_layer):
                model.add(Dense(dense_layer_size))
                model.add(Activation("relu"))
                model.add(Dropout(0.2))

            model.add(Dense(1))
            model.add(Activation("sigmoid"))

            model.compile(loss="binary_crossentropy", 
                          optimizer="adam", 
                          metrics=["accuracy"])

            model.fit(x, y, batch_size=32, epochs=7, validation_split=0.15, callbacks=[tensorboard]) # 7 epochs appears best for 0-dense-3-conv-96-nodes
            # Dropout doesn't appear to make any difference with one 512 Dense layer added.

model.save("96x3-CNN.model")

Note the model.save("96x3-CNN.model"). We can now use this in a completely separate file to test an image:

import cv2
import tensorflow as tf


CATEGORIES = ['Dog', 'Cat']

def prepare(filepath):
    IMG_SIZE = 50
    img_array = cv2.imread(filepath, cv2.IMREAD_GRAYSCALE)
    img_array = img_array/255.0
    new_array = cv2.resize(img_array, (IMG_SIZE, IMG_SIZE))
    return new_array.reshape(-1, IMG_SIZE, IMG_SIZE, 1)

model = tf.keras.models.load_model("96x3-CNN.model")
prediction = model.predict([prepare("dog.jpg")])
print(CATEGORIES[int(round(prediction[0][0]))])

This outputs "Dog". The dog.jpg file is sitting in the same directory as all these other files. We are just using CATEGORIES here to output a readable string. The prediction variable will actually output a number between 0 and 1, which gives some indication of how certain the model is that the image is a cat or dog. I'm then using round and int to call the appropriate index in the CATEGORIES list.