Writing a training loop from scratch in TensorFlow

import time
import os

# This guide can only be run with the TensorFlow backend.
os.environ["KERAS_BACKEND"] = "tensorflow"

import tensorflow as tf
import keras
import numpy as np

def get_model():
    inputs = keras.Input(shape=(784,), name="digits")
    x1 = keras.layers.Dense(64, activation="relu")(inputs)
    x2 = keras.layers.Dense(64, activation="relu")(x1)
    outputs = keras.layers.Dense(10, name="predictions")(x2)
    model = keras.Model(inputs=inputs, outputs=outputs)
    return model


model = get_model()

# Instantiate an optimizer.
optimizer = keras.optimizers.Adam(learning_rate=1e-3)
# Instantiate a loss function.
loss_fn = keras.losses.SparseCategoricalCrossentropy(from_logits=True)

# Prepare the training dataset.
batch_size = 32
(x_train, y_train), (x_test, y_test) = keras.datasets.mnist.load_data()
x_train = np.reshape(x_train, (-1, 784))
x_test = np.reshape(x_test, (-1, 784))

# Reserve 10,000 samples for validation.
x_val = x_train[-10000:]
y_val = y_train[-10000:]
x_train = x_train[:-10000]
y_train = y_train[:-10000]

# Prepare the training dataset.
train_dataset = tf.data.Dataset.from_tensor_slices((x_train, y_train))
train_dataset = train_dataset.shuffle(buffer_size=1024).batch(batch_size)

# Prepare the validation dataset.
val_dataset = tf.data.Dataset.from_tensor_slices((x_val, y_val))
val_dataset = val_dataset.batch(batch_size)

epochs = 3
for epoch in range(epochs):
    print(f"\nStart of epoch {epoch}")

    # Iterate over the batches of the dataset.
    for step, (x_batch_train, y_batch_train) in enumerate(train_dataset):
        # Open a GradientTape to record the operations run
        # during the forward pass, which enables auto-differentiation.
        with tf.GradientTape() as tape:
            # Run the forward pass of the layer.
            # The operations that the layer applies
            # to its inputs are going to be recorded
            # on the GradientTape.
            logits = model(x_batch_train, training=True)  # Logits for this minibatch

            # Compute the loss value for this minibatch.
            loss_value = loss_fn(y_batch_train, logits)

        # Use the gradient tape to automatically retrieve
        # the gradients of the trainable variables with respect to the loss.
        grads = tape.gradient(loss_value, model.trainable_weights)

        # Run one step of gradient descent by updating
        # the value of the variables to minimize the loss.
        optimizer.apply(grads, model.trainable_weights)

        # Log every 100 batches.
        if step % 100 == 0:
            print(
                f"Training loss (for 1 batch) at step {step}: {float(loss_value):.4f}"
            )
            print(f"Seen so far: {(step + 1) * batch_size} samples")

# Get a fresh model
model = get_model()

# Instantiate an optimizer to train the model.
optimizer = keras.optimizers.Adam(learning_rate=1e-3)
# Instantiate a loss function.
loss_fn = keras.losses.SparseCategoricalCrossentropy(from_logits=True)

# Prepare the metrics.
train_acc_metric = keras.metrics.SparseCategoricalAccuracy()
val_acc_metric = keras.metrics.SparseCategoricalAccuracy()

epochs = 2
for epoch in range(epochs):
    print(f"\nStart of epoch {epoch}")
    start_time = time.time()

    # Iterate over the batches of the dataset.
    for step, (x_batch_train, y_batch_train) in enumerate(train_dataset):
        with tf.GradientTape() as tape:
            logits = model(x_batch_train, training=True)
            loss_value = loss_fn(y_batch_train, logits)
        grads = tape.gradient(loss_value, model.trainable_weights)
        optimizer.apply(grads, model.trainable_weights)

        # Update training metric.
        train_acc_metric.update_state(y_batch_train, logits)

        # Log every 100 batches.
        if step % 100 == 0:
            print(
                f"Training loss (for 1 batch) at step {step}: {float(loss_value):.4f}"
            )
            print(f"Seen so far: {(step + 1) * batch_size} samples")

    # Display metrics at the end of each epoch.
    train_acc = train_acc_metric.result()
    print(f"Training acc over epoch: {float(train_acc):.4f}")

    # Reset training metrics at the end of each epoch
    train_acc_metric.reset_state()

    # Run a validation loop at the end of each epoch.
    for x_batch_val, y_batch_val in val_dataset:
        val_logits = model(x_batch_val, training=False)
        # Update val metrics
        val_acc_metric.update_state(y_batch_val, val_logits)
    val_acc = val_acc_metric.result()
    val_acc_metric.reset_state()
    print(f"Validation acc: {float(val_acc):.4f}")
    print(f"Time taken: {time.time() - start_time:.2f}s")

@tf.function
def train_step(x, y):
    with tf.GradientTape() as tape:
        logits = model(x, training=True)
        loss_value = loss_fn(y, logits)
    grads = tape.gradient(loss_value, model.trainable_weights)
    optimizer.apply(grads, model.trainable_weights)
    train_acc_metric.update_state(y, logits)
    return loss_value

@tf.function
def test_step(x, y):
    val_logits = model(x, training=False)
    val_acc_metric.update_state(y, val_logits)

epochs = 2
for epoch in range(epochs):
    print(f"\nStart of epoch {epoch}")
    start_time = time.time()

    # Iterate over the batches of the dataset.
    for step, (x_batch_train, y_batch_train) in enumerate(train_dataset):
        loss_value = train_step(x_batch_train, y_batch_train)

        # Log every 100 batches.
        if step % 100 == 0:
            print(
                f"Training loss (for 1 batch) at step {step}: {float(loss_value):.4f}"
            )
            print(f"Seen so far: {(step + 1) * batch_size} samples")

    # Display metrics at the end of each epoch.
    train_acc = train_acc_metric.result()
    print(f"Training acc over epoch: {float(train_acc):.4f}")

    # Reset training metrics at the end of each epoch
    train_acc_metric.reset_state()

    # Run a validation loop at the end of each epoch.
    for x_batch_val, y_batch_val in val_dataset:
        test_step(x_batch_val, y_batch_val)

    val_acc = val_acc_metric.result()
    val_acc_metric.reset_state()
    print(f"Validation acc: {float(val_acc):.4f}")
    print(f"Time taken: {time.time() - start_time:.2f}s")

class ActivityRegularizationLayer(keras.layers.Layer):
    def call(self, inputs):
        self.add_loss(1e-2 * tf.reduce_sum(inputs))
        return inputs

inputs = keras.Input(shape=(784,), name="digits")
x = keras.layers.Dense(64, activation="relu")(inputs)
# Insert activity regularization as a layer
x = ActivityRegularizationLayer()(x)
x = keras.layers.Dense(64, activation="relu")(x)
outputs = keras.layers.Dense(10, name="predictions")(x)

model = keras.Model(inputs=inputs, outputs=outputs)

@tf.function
def train_step(x, y):
    with tf.GradientTape() as tape:
        logits = model(x, training=True)
        loss_value = loss_fn(y, logits)
        # Add any extra losses created during the forward pass.
        loss_value += sum(model.losses)
    grads = tape.gradient(loss_value, model.trainable_weights)
    optimizer.apply(grads, model.trainable_weights)
    train_acc_metric.update_state(y, logits)
    return loss_value

discriminator = keras.Sequential(
    [
        keras.Input(shape=(28, 28, 1)),
        keras.layers.Conv2D(64, (3, 3), strides=(2, 2), padding="same"),
        keras.layers.LeakyReLU(negative_slope=0.2),
        keras.layers.Conv2D(128, (3, 3), strides=(2, 2), padding="same"),
        keras.layers.LeakyReLU(negative_slope=0.2),
        keras.layers.GlobalMaxPooling2D(),
        keras.layers.Dense(1),
    ],
    name="discriminator",
)
discriminator.summary()

latent_dim = 128

generator = keras.Sequential(
    [
        keras.Input(shape=(latent_dim,)),
        # We want to generate 128 coefficients to reshape into a 7x7x128 map
        keras.layers.Dense(7 * 7 * 128),
        keras.layers.LeakyReLU(negative_slope=0.2),
        keras.layers.Reshape((7, 7, 128)),
        keras.layers.Conv2DTranspose(128, (4, 4), strides=(2, 2), padding="same"),
        keras.layers.LeakyReLU(negative_slope=0.2),
        keras.layers.Conv2DTranspose(128, (4, 4), strides=(2, 2), padding="same"),
        keras.layers.LeakyReLU(negative_slope=0.2),
        keras.layers.Conv2D(1, (7, 7), padding="same", activation="sigmoid"),
    ],
    name="generator",
)

# Instantiate one optimizer for the discriminator and another for the generator.
d_optimizer = keras.optimizers.Adam(learning_rate=0.0003)
g_optimizer = keras.optimizers.Adam(learning_rate=0.0004)

# Instantiate a loss function.
loss_fn = keras.losses.BinaryCrossentropy(from_logits=True)


@tf.function
def train_step(real_images):
    # Sample random points in the latent space
    random_latent_vectors = tf.random.normal(shape=(batch_size, latent_dim))
    # Decode them to fake images
    generated_images = generator(random_latent_vectors)
    # Combine them with real images
    combined_images = tf.concat([generated_images, real_images], axis=0)

    # Assemble labels discriminating real from fake images
    labels = tf.concat(
        [tf.ones((batch_size, 1)), tf.zeros((real_images.shape[0], 1))], axis=0
    )
    # Add random noise to the labels - important trick!
    labels += 0.05 * tf.random.uniform(labels.shape)

    # Train the discriminator
    with tf.GradientTape() as tape:
        predictions = discriminator(combined_images)
        d_loss = loss_fn(labels, predictions)
    grads = tape.gradient(d_loss, discriminator.trainable_weights)
    d_optimizer.apply(grads, discriminator.trainable_weights)

    # Sample random points in the latent space
    random_latent_vectors = tf.random.normal(shape=(batch_size, latent_dim))
    # Assemble labels that say "all real images"
    misleading_labels = tf.zeros((batch_size, 1))

    # Train the generator (note that we should *not* update the weights
    # of the discriminator)!
    with tf.GradientTape() as tape:
        predictions = discriminator(generator(random_latent_vectors))
        g_loss = loss_fn(misleading_labels, predictions)
    grads = tape.gradient(g_loss, generator.trainable_weights)
    g_optimizer.apply(grads, generator.trainable_weights)
    return d_loss, g_loss, generated_images

# Prepare the dataset. We use both the training & test MNIST digits.
batch_size = 64
(x_train, _), (x_test, _) = keras.datasets.mnist.load_data()
all_digits = np.concatenate([x_train, x_test])
all_digits = all_digits.astype("float32") / 255.0
all_digits = np.reshape(all_digits, (-1, 28, 28, 1))
dataset = tf.data.Dataset.from_tensor_slices(all_digits)
dataset = dataset.shuffle(buffer_size=1024).batch(batch_size)

epochs = 1  # In practice you need at least 20 epochs to generate nice digits.
save_dir = "./"

for epoch in range(epochs):
    print(f"\nStart epoch {epoch}")

    for step, real_images in enumerate(dataset):
        # Train the discriminator & generator on one batch of real images.
        d_loss, g_loss, generated_images = train_step(real_images)

        # Logging.
        if step % 100 == 0:
            # Print metrics
            print(f"discriminator loss at step {step}: {d_loss:.2f}")
            print(f"adversarial loss at step {step}: {g_loss:.2f}")

            # Save one generated image
            img = keras.utils.array_to_img(generated_images[0] * 255.0, scale=False)
            img.save(os.path.join(save_dir, f"generated_img_{step}.png"))

        # To limit execution time we stop after 10 steps.
        # Remove the lines below to actually train the model!
        if step > 10:
            break