Closed Andyloris closed 2 years ago
Going to need more information. Do you have some sample code you're working with?
This is an example of a decision making function.. You want to return format_decision(self, <move_choice>)
. Something like : format_decision(self, "flamethrower")
, or format_decision(self, "switch landorustherian")
I need to locally run a match beetween the bot and itself. How could i do this ?
I need to locally run a match beetween the bot and itself. How could i do this ?
Two ways you can do this:
For option number 2 you want to generate instructions with:
instructions = get_all_state_instructions(mutator, my_move, your_move)
Pick a random instruction set based on the chances of each happening
chosen_instructions = random.choices(instructions, weights=[i.percentage for i in instructions])
and then apply one of the instructions with:
mutator.apply(chosen_instructions)
I think that option 1 is faster than option 2 if you disable throttling, which can be achieved with the --no-security
flag - I got up to 20 games per second with it.
I think that option 1 is faster than option 2 if you disable throttling, which can be achieved with the
--no-security
flag - I got up to 20 games per second with it.
The bottleneck running this project's bot against itself will surely be running minimax + doing evaluations, not the PS server. Especially if something like depth=2 is used. Even at depth=1 games will still take ~40 seconds each. Running multiple sessions of the bot vs itself will speed it up, but you'll need the compute resources to keep up.
Hi, i'm trying to implement tuning in showdown and, how can i play a move in the state class ?