tf.mul, and tf.sub are deprecated in favor of tf.multiply and tf.subtract I fixed with: sed -i 's/tf.mul/tf.multiply/g' dqn.py sed -i 's/tf.sub/tf.subtract/g' dqn.py