class ActionGiver():
def __init__(self, args):
arg_parser = build_arg_parser(args)
self.world = RLWorld(arg_parser)
def get_ac(self, s, g):
return self.world.get_action(s, g)
def build_world(args, enable_draw, playback_speed=1):
arg_parser = build_arg_parser(args)
# build env: core, seed
env = DeepMimicEnv(args, enable_draw)
# build world for agent
world = RLWorld(env, arg_parser)
world.env.set_playback_speed(playback_speed)
return world
def build_world(args, enable_draw, playback_speed=1):
arg_parser = build_arg_parser(args)
env = DeepMimicEnv(args, enable_draw)
world = RLWorld(env, arg_parser)
world.env.set_playback_speed(playback_speed)
return world
def __init__(self, args):
arg_parser = build_arg_parser(args)
self.world = RLWorld(arg_parser)
enable_draw = True
timestep = 1. / 240.
args = sys.argv[1:]
arg_parser = ArgParser()
arg_parser.load_args(args)
arg_file = arg_parser.parse_string('arg_file',
"run_humanoid3d_spinkick_args.txt")
arg_parser.load_file(pybullet_data.getDataPath() + "/args/" + arg_file)
motion_file = arg_parser.parse_strings('motion_file')
fall_contact_bodies = arg_parser.parse_ints("fall_contact_bodies")
env = PyBulletDeepMimicEnv(motion_file, enable_draw, fall_contact_bodies)
world = RLWorld(env, arg_parser)
world.reset()
total_reward = 0
steps = 0
while True:
world.update(timestep)
total_reward += world.env.calc_reward(agent_id=0)
steps += 1
if world.env.is_episode_end() or steps >= 1000:
total_reward = 0