Pruning, Weight Sharing and Quantization are commonly used techniques and can be done in tf.