Hello Tyun,

Thank you for interests in my post. Unfortunately, the video of this post is recorded using the retrained model that is provided by the OpenAI.

I also could not find the training source code. However, I am going to implement it soon myself for next post.