def __init__(self):
self.logger = utils.getLogger()
self.num_genomes = 0
"""metrics to compare different containers"""
self.max_quality = -np.inf
self.min_quality = np.inf
self.total_quality = 0
def __init__(self,env,genome_constructor,batch_size,seed=1): self.env = env self.genome_constructor = genome_constructor self.batch_size = batch_size self.seed = seed self.logger = utils.getLogger() self.current_iteration = 1
def __init__(self, parameters=None, seed=1):
self.logger = utils.getLogger()
if (parameters == None):
self.parameters = self.sample_random_genome()
else:
self.parameters = parameters
self.seed = seed
## Set random number seed for all scipy and numpy operations so that experiments can be reproduced
np.random.seed(seed=seed)
from mpi4py import MPI
from evo_rbc.env.prosthetic_env import ProstheticEAEnv
import evo_rbc.main.utils as test_utils
logger = test_utils.getLogger()
comm = MPI.Comm.Get_parent()
size = comm.Get_size()
rank = comm.Get_rank()
genomes_matrix = visualise = max_time_steps_qd = joint_error_margin = None
max_time_steps_qd = comm.bcast(max_time_steps_qd, root=0)
# print("child spawned and running",rank)
genomes = comm.scatter(genomes_matrix, root=0)
visualise = comm.bcast(visualise, root=0)
joint_error_margin = comm.bcast(joint_error_margin, root=0)
env = ProstheticEAEnv(max_time_steps_qd=max_time_steps_qd,
joint_error_margin=joint_error_margin)
qd_function = env.qd_steady_runner
# print("visualise",visualise,rank)
qd_evaluations = []
for i in range(len(genomes)):
behavior, quality = env.evaluate_quality_diversity_fitness(
qd_function=qd_function,
def __setstate__(self, state):
for k, v in state.items():
self.__dict__[k] = v
self.logger = utils.getLogger()
def __init__(self):
self.logger = utils.getLogger()