pip3 install -r requirement.txtpython3 main.py [team_name_1] [team_name_2]usage: main.py [-h] [--server_port SERVER_PORT] [-f1 FILE1] [-f2 FILE2] [-o]
[--player_num_1 PLAYER_NUM_1] [--player_num_2 PLAYER_NUM_2]
[-l] [-s] [-w] [--gamma GAMMA] [--learning_rate LEARNING_RATE]
[--save_interval SAVE_INTERVAL] [--address ADDRESS]
[--host HOST] [--send_log] [--recieve_log]
[--analysis_log [ANALYSIS_LOG [ANALYSIS_LOG ...]]]
[--logdir LOGDIR] [--max_ball_distance MAX_BALL_DISTANCE]
[--min_ball_distance MIN_BALL_DISTANCE]
[--ball_distance_interval BALL_DISTANCE_INTERVAL]
[--special_division]
[--special_division_value SPECIAL_DIVISION_VALUE]
[--damage_reset_ball_distance DAMAGE_RESET_BALL_DISTANCE]
[--max_ball_angle MAX_BALL_ANGLE]
[--min_ball_angle MIN_BALL_ANGLE]
[--ball_angle_interval BALL_ANGLE_INTERVAL]
[--max_goal_distance MAX_GOAL_DISTANCE]
[--min_goal_distance MIN_GOAL_DISTANCE]
[--goal_distance_interval GOAL_DISTANCE_INTERVAL]
[--max_goal_angle MAX_GOAL_ANGLE]
[--min_goal_angle MIN_GOAL_ANGLE]
[--goal_angle_interval GOAL_ANGLE_INTERVAL] [--noise NOISE]
[--actions ACTIONS] [--goal_reset GOAL_RESET]
team_name_1 team_name_2
A program for learning agent in a robocup 2D simulation league
positional arguments:
team_name_1 Specify a team name for Team 1
team_name_2 Specify a team name for Team 2
optional arguments:
-h, --help show this help message and exit
--server_port SERVER_PORT
Specifies the server port to connect to. If you
specify --auto_start argument, the server port will be
set to this value, and the It is.
-f1 FILE1, --file1 FILE1
Specify the trained data for team 1
-f2 FILE2, --file2 FILE2
Specify the trained data for team 2
-o, --only_team_1 Register one team only
--player_num_1 PLAYER_NUM_1
Set the number of players registered for Team 1
--player_num_2 PLAYER_NUM_2
Set the number of players registered for Team 2
-l, --logging Save the trained data
-s, --auto_server Auto-start the server.
-w, --auto_window Auto-start soccerwindow2 (must be installed
beforehand).
--gamma GAMMA Specifies the gamma of the ActorCritic.train function
--learning_rate LEARNING_RATE
Specify the learning rate of soccerai
--save_interval SAVE_INTERVAL
Specify the storage interval of learned data in
episode.
--address ADDRESS The IP address of the server to connect to
--host HOST Enter the host's address
--send_log Display the logs you sent to the server
--recieve_log Display the logs received from the server
--analysis_log [ANALYSIS_LOG [ANALYSIS_LOG ...]]
Display the log of messages received from the server
(specify multiple messages in the filter)
--logdir LOGDIR Specify the location of the tensorboard's log file
(the data will be deleted if the specified location
exists)
--max_ball_distance MAX_BALL_DISTANCE
Specifies the maximum distance of the ball that the AI
can recognize. If the maximum value is exceeded, it
will be cut down to the maximum value.
--min_ball_distance MIN_BALL_DISTANCE
Specifies the minimum distance of the ball that the AI
can recognize.
--ball_distance_interval BALL_DISTANCE_INTERVAL
The distance specifies how ambiguous the number
between the distances of the ball is to be learned.
--special_division Create a dedicated Q table index when a ball is seen
in a specific range
--special_division_value SPECIAL_DIVISION_VALUE
Specify a specific range of special_division
--damage_reset_ball_distance DAMAGE_RESET_BALL_DISTANCE
Specify the distance of the ball as a condition for
resetting the damage you receive per episode.
--max_ball_angle MAX_BALL_ANGLE
Specifies the maximum angle relative to the ball that
the AI can recognize. If the relative angle is greater
than the maximum value, it will be cut down to the
maximum value.
--min_ball_angle MIN_BALL_ANGLE
Specifies the minimum relative angle to the ball that
the AI can recognize. If the relative angle is smaller
than the minimum value, it is rounded up to the
minimum value.
--ball_angle_interval BALL_ANGLE_INTERVAL
The angle specifies how ambiguous the relative angle
of the ball is to be learned.
--max_goal_distance MAX_GOAL_DISTANCE
Specifies the maximum angle relative to the goal that
the AI can recognize. If the relative angle is greater
than the maximum value, it will be cut down to the
maximum value.
--min_goal_distance MIN_GOAL_DISTANCE
Specifies the minimum relative angle to the goal that
the AI can recognize. If the relative angle is smaller
than the minimum value, it is rounded up to the
minimum value.
--goal_distance_interval GOAL_DISTANCE_INTERVAL
The angle specifies how ambiguous the relative angle
of the goal is to be learned.
--max_goal_angle MAX_GOAL_ANGLE
Specifies the maximum angle relative to the goal that
the AI can recognize. If the relative angle is greater
than the maximum value, it will be cut down to the
maximum value.
--min_goal_angle MIN_GOAL_ANGLE
Specifies the minimum relative angle to the goal that
the AI can recognize. If the relative angle is smaller
than the minimum value, it is rounded up to the
minimum value.
--goal_angle_interval GOAL_ANGLE_INTERVAL
The angle specifies how ambiguous the relative angle
of the goal is to be learned.
--noise NOISE How much of the actions given by the AI will be used
--actions ACTIONS Specify the number of actions
--goal_reset GOAL_RESET
Specify the specific episode in the graph that shows
the number of times a goal is scored within a specific
episode displayed on tensorboard.