dqn_agent_nature.py 文件源码

def __init__(self, gpu_id, state_dimention,batchsize,historysize, enable_controller,arch):
        self.gpu_id = gpu_id
        self.num_of_actions = len(enable_controller)
        self.enable_controller = enable_controller  # Default setting : "Pong"
        self.arch = arch
        self.replay_size = batchsize
        self.data_size = historysize

        self.state_dimention = state_dimention
        print "Initializing DQN..."
        #   Initialization of Chainer 1.1.0 or older.
        #        print "CUDA init"
        #        cuda.init()

        print "Model Building"
        #self.model = dnn_6_new.Q_DNN(self.state_dimention,200,self.num_of_actions)
        self.model = self.set_model(self.arch,self.state_dimention,200,self.num_of_actions)
        self.model.to_gpu(self.gpu_id)


        self.model_target = copy.deepcopy(self.model)

        print "Initizlizing Optimizer"
        self.optimizer = optimizers.RMSpropGraves(lr=0.00025, alpha=0.95, momentum=0.95, eps=0.0001)
        self.optimizer.setup(self.model)

        # History Data :  D=[s, a, r, s_dash, end_episode_flag]
        self.D = [np.zeros((self.data_size, 1, self.state_dimention), dtype=np.float32),
                  np.zeros(self.data_size, dtype=np.int8),
                  np.zeros((self.data_size, 1), dtype=np.float32),
                  np.zeros((self.data_size, 1, self.state_dimention), dtype=np.float32),
                  np.zeros((self.data_size, 1), dtype=np.bool)]