Keras + generator, Minimal example

minimal keras training from generator example

import numpy as np
import tensorflow as tf
from tensorflow.keras import Sequential
from tensorflow.keras.layers import Dense

batch_size = 1
data_size = 1


def generator():
    print('initializing generator')
    values = list(range(10))
    for v in values:
        print(f'"{v}"')
        data = np.ones((batch_size, data_size)) * v
        targets = np.random.random((batch_size, 1))
        yield (data, targets)
    print('no more values left')

model = Sequential([Dense(10), Dense(data_size)])
model.compile(optimizer='adam', loss='mae')
model.build((batch_size, data_size))
model.summary()

output_types = (tf.float64, tf.float64)
output_shapes = (tf.TensorShape((batch_size, data_size)),
                 tf.TensorShape((batch_size, 1)))
dataset = tf.data.Dataset.from_generator(generator, output_types=output_types, output_shapes=output_shapes)
dataset = dataset.repeat()

model.fit(dataset, epochs=2, steps_per_epoch=6)