2023-03-23 Performance of prediction in this post is not practical, a single token per a minute. The next post shows the other case. Its performance is better due to small size language model. If you are interested, please refer to it. imp…