def _define_simulated_batch_env(hparams,
num_agents,
simulation_random_starts=False,
intrinsic_reward_scale=0.):
cur_batch_env = simulated_batch_env.SimulatedBatchEnv(
hparams, num_agents, simulation_random_starts, intrinsic_reward_scale)
return cur_batch_env
def define_simulated_batch_env(num_agents):
# TODO(blazej0): the parameters should be infered.
observ_shape = (210, 160, 3)
observ_dtype = tf.float32
action_shape = []
action_dtype = tf.int32
cur_batch_env = simulated_batch_env.SimulatedBatchEnv(
num_agents, observ_shape, observ_dtype, action_shape, action_dtype)
return cur_batch_env
def define_simulated_batch_env(environment_lambda,
num_agents,
problem,
simulation_random_starts=False,
intrinsic_reward_scale=0.):
cur_batch_env = simulated_batch_env.SimulatedBatchEnv(
environment_lambda, num_agents, problem, simulation_random_starts,
intrinsic_reward_scale)
return cur_batch_env
def define_simulated_batch_env(num_agents):
#TODO: pm->Błażej. Should the paramters be infered.
observ_shape, observ_dtype, action_shape, action_dtype = (
210, 160, 3), tf.float32, [], tf.int32
batch_env = simulated_batch_env.SimulatedBatchEnv(num_agents, observ_shape,
observ_dtype,
action_shape,
action_dtype)
return batch_env
def define_simulated_batch_env(environment_lambda, num_agents, problem):
cur_batch_env = simulated_batch_env.SimulatedBatchEnv(
environment_lambda, num_agents, problem)
return cur_batch_env
def _define_simulated_batch_env(environment_spec, num_agents,
initial_frame_chooser):
cur_batch_env = simulated_batch_env.SimulatedBatchEnv(
environment_spec, num_agents, initial_frame_chooser)
return cur_batch_env
def _define_simulated_batch_env(environment_spec, num_agents):
cur_batch_env = simulated_batch_env.SimulatedBatchEnv(environment_spec,
num_agents)
return cur_batch_env
