/keras-tpu

A demo to show how to use the TPU in google colab

Primary LanguageJupyter Notebook

keras tpu

A demo to show how to use the TPU in google colab.

It is almost the same as training a normal keras model, except that you need to use tf.contrib.tpu.keras_to_tpu_model to transfer the model to TPU:

tpu_model = tf.contrib.tpu.keras_to_tpu_model(
    model,
    strategy=tf.contrib.tpu.TPUDistributionStrategy(
        tf.contrib.cluster_resolver.TPUClusterResolver(tpu=TPU_ADDRESS)
    )
)
tpu_model.compile(
    optimizer=tf.train.AdamOptimizer(learning_rate=1e-3, ),
    loss=tf.keras.losses.sparse_categorical_crossentropy,
    metrics=['sparse_categorical_accuracy']
)

You can find more info in the code or checkout the colab notebook