I'm doing a job for an associate who runs a poker school. I am taking over for him as he is not nearly as technically inclined as I am however AI falls outside my scope of knowledge and this maybe just a little too complicated for me. What I think I have is a fully working trainer for what is essentially a copy of Pluribus facebooks 6 player no limit texas holdem AI. However I would like someone to validate that this is the case and fine tune the variables that need to set such as the the final compression and length of run time ect. Further and most importantly I would like someone to implement the realtime search such that you can either play against the ai (5 AI vs 1 human) or so that the ai can play against a table of humans (1AI 5 humans) for both cases I would like to be able to be able to edit the information that is important to the ai things such as where the dealer button is the players stacks (chips) the pot and any other information that maybe needed for the ai to make its next move. Finally for the second version (1AI 5 humans) I would like to be able to run multiple instances of the AI at one time.
If the research by facebook is to be believed I have the hardware necessary to train the AI. A a dual 32 core epyc system with 512gb of ram. This should be sufficient. If you require any of the original papers I can get them for you.
7 freelances font une offre moyenne de 50 $/heure pour ce travail
Hello, there. I worked on several AI games required reinforcement learning, so can definitely work on this project. Please get in touch with me so we can discuss more. Thank you.