import tensorflow as tf

# Convert the model
saved_model_dir = ''
converter = tf.lite.TFLiteConverter.from_saved_model(
    saved_model_dir)  # path to the SavedModel directory
tflite_model = converter.convert()

# Save the model.
with open('model.tflite', 'wb') as f:
    f.write(tflite_model)