args = parser.parse_args()
#-----Environment-----#
width = height = 12
obstacles = create_obstacles(width, height)
set_diff = list(set(product(tuple(range(3, width-3)), repeat=2)) - set(obstacles))
start_loc = sample_start(set_diff)
s = State(start_loc, obstacles)
T = TransitionFunction(width, height, obstacle_movement)
if args.expert_path == 'SR2_expert_trajectories/':
R = RewardFunction_SR2(-1.0,1.0,width)
else:
R = RewardFunction(-1.0,1.0)
num_inputs = s.state.shape[0]
num_actions = 4
if args.expert_path == 'SR2_expert_trajectories/':
num_c = 2
else:
num_c = 4
#env.seed(args.seed)
torch.manual_seed(args.seed)
policy_net = GRU(num_inputs,
0,
num_c,
num_actions,
args.cuda = not args.no_cuda and torch.cuda.is_available()
torch.manual_seed(args.seed)
if args.cuda:
torch.cuda.manual_seed(args.seed)
#-----Environment-----#
width = height = 12
obstacles = create_obstacles(width, height)
T = TransitionFunction(width, height, obstacle_movement)
if args.expert_path == 'SR2_expert_trajectories/':
R = RewardFunction_SR2(-1.0, 1.0, width)
else:
R = RewardFunction(-1.0, 1.0)
class VAE(nn.Module):
def __init__(self):
super(VAE, self).__init__()
self.policy = Policy(state_size=8,
action_size=0,
latent_size=2,
output_size=4,
hidden_size=64,
output_activation='sigmoid')
self.posterior = Posterior(state_size=8,
action_size=0,
latent_size=2,