OpenAI Five Benchmark

Re:OpenAI Five Benchmark by lightbyte on Wednesday July 18, 2018 @01:21PM (#17560254)

I think OP wants to know how the neural network actually plays the game. I think in this case the dota client has an api for bots that it can use?

[ Parent | on HN ]

Re:OpenAI Five Benchmark by gdb on Wednesday July 18, 2018 @01:30PM (#17560324)

Yes, there's a bot API.

We dump state from the bot API each tick and send it over GRPC to a Python agent, which formats the state into a tuple of Numpy arrays. That Numpy array is passed into 5 neural networks (one per agent), each of which returns a tuple of Numpy arrays. Each tuple is decoded into a semantic action, which is then returned to the game via GRPC.

[ Parent | on HN ]

Re:OpenAI Five Benchmark by chlvsl on Wednesday July 18, 2018 @03:16PM (#17561280)

Is the entire system one agent, which is then replicated across 5 bot instances, or do you have a specific network per hero?

[ Parent | on HN ]