Python chainer.functions.Linear() Examples

The following are 12 code examples of chainer.functions.Linear(). You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may also want to check out all available functions/classes of the module chainer.functions , or try the search function .
Example #1
Source File: run_tiger.py    From Chimp with Apache License 2.0 5 votes vote down vote up 
def __init__(self):
        super(Linear, self).__init__(
            l1=F.Bilinear(settings["n_frames"], settings["n_frames"], 200),
            l2=F.Linear(200, 100, wscale=np.sqrt(2)),
            l3=F.Linear(100, 100, wscale=np.sqrt(2)),
            l4=F.Linear(100, 50, wscale=np.sqrt(2)),
            l5=F.Linear(50, simulator.n_actions, wscale = np.sqrt(2))
        )
Example #2
Source File: mountain_car_test.py    From Chimp with Apache License 2.0 5 votes vote down vote up 
def __init__(self):
        super(TestNet, self).__init__(
            l1=F.Linear(settings['model_dims'][1], 20, bias=0.0),
            l2=F.Linear(20, 10, bias=0.0),
            bn1=L.BatchNormalization(10),
            l3=F.Linear(10, 10),
            l4=F.Linear(10, 10),
            bn2=L.BatchNormalization(10),
            lout=F.Linear(10, simulator.n_actions)
        )
        self.train = True
        # initialize avg_var to prevent divide by zero
        self.bn1.avg_var.fill(0.1),
        self.bn2.avg_var.fill(0.1),
Example #3
Source File: run_atari.py    From Chimp with Apache License 2.0 5 votes vote down vote up 
def __init__(self):
        super(Convolution, self).__init__(
            l1=F.Convolution2D(settings['history_sizes'][0], 32, ksize=8, stride=4, nobias=False, wscale=np.sqrt(2)),
            l2=F.Convolution2D(32, 64, ksize=4, stride=2, nobias=False, wscale=np.sqrt(2)),
            l3=F.Convolution2D(64, 64, ksize=3, stride=1, nobias=False, wscale=np.sqrt(2)),
            l4=F.Linear(3136, 512, wscale = np.sqrt(2)),
            l5=F.Linear(512, simulator.n_actions, wscale = np.sqrt(2)),
        )
Example #4
Source File: run_mountain_car.py    From Chimp with Apache License 2.0 5 votes vote down vote up 
def __init__(self):
        super(CarNet, self).__init__(
            l1=F.Linear(settings['model_dims'][1], 20, bias=0.0),
            l2=F.Linear(20, 10, bias=0.0),
            bn1=L.BatchNormalization(10),
            l3=F.Linear(10, 10),
            l4=F.Linear(10, 10),
            bn2=L.BatchNormalization(10),
            lout=F.Linear(10, simulator.n_actions)
        )
        self.train = True
        # initialize avg_var to prevent divide by zero
        self.bn1.avg_var.fill(0.1),
        self.bn2.avg_var.fill(0.1),
Example #5
Source File: run_cartpole.py    From Chimp with Apache License 2.0 5 votes vote down vote up 
def __init__(self):
        super(CartNet, self).__init__(
            l1=F.Linear(4, 20, bias=0.0),
            l2=F.Linear(20, 10, bias=0.0),
            bn1=L.BatchNormalization(10),
            l3=F.Linear(10, 10),
            l4=F.Linear(10, 10),
            bn2=L.BatchNormalization(10),
            lout=F.Linear(10, simulator.n_actions)
        )
        self.train = True
        # initialize avg_var to prevent divide by zero
        self.bn1.avg_var.fill(0.1),
        self.bn2.avg_var.fill(0.1),
Example #6
Source File: agent_test.py    From Chimp with Apache License 2.0 5 votes vote down vote up 
def __init__(self):
        super(TestNet, self).__init__(
            l1=F.Linear(settings['model_dims'][1], 20, bias=0.0),
            l2=F.Linear(20, 10, bias=0.0),
            bn1=L.BatchNormalization(10),
            l3=F.Linear(10, 10),
            l4=F.Linear(10, 10),
            bn2=L.BatchNormalization(10),
            lout=F.Linear(10, simulator.n_actions)
        )
        self.train = True
        # initialize avg_var to prevent divide by zero
        self.bn1.avg_var.fill(0.1),
        self.bn2.avg_var.fill(0.1),
Example #7
Source File: confirm_dqn_env.py    From techcircle_openai_handson with MIT License 5 votes vote down vote up 
def __init__(self, action, other_action, size, epsilon=0.05, hidden=200):
        self.action = action
        self.other_action = other_action
        self.width = size * size
        self.epsilon = epsilon
        self.hidden = hidden
        super(ChainerAgent, self).__init__(
            l1=F.Linear(self.width, self.hidden, wscale=np.sqrt(2)),
            l2=F.Linear(self.hidden, 1, wscale=np.sqrt(2)),
        )
Example #8
Source File: dqn_agent.py    From DQN-chainer with MIT License 5 votes vote down vote up 
def __init__(self, n_history, n_act):
        super(ActionValue, self).__init__(
            l1=F.Convolution2D(n_history, 32, ksize=8, stride=4, nobias=False, wscale=np.sqrt(2)),
            l2=F.Convolution2D(32, 64, ksize=4, stride=2, nobias=False, wscale=np.sqrt(2)),
            l3=F.Convolution2D(64, 64, ksize=3, stride=1, nobias=False, wscale=np.sqrt(2)),
            l4=F.Linear(3136, 512, wscale=np.sqrt(2)),
            q_value=F.Linear(512, n_act,
                             initialW=np.zeros((n_act, 512),
                             dtype=np.float32))
        )
Example #9
Source File: dqn_agent_cpu.py    From DQN-chainer with MIT License 5 votes vote down vote up 
def __init__(self, n_history, n_act):
        super(ActionValue, self).__init__(
            l1=F.Convolution2D(n_history, 32, ksize=8, stride=4, nobias=False, wscale=np.sqrt(2)),
            l2=F.Convolution2D(32, 64, ksize=4, stride=2, nobias=False, wscale=np.sqrt(2)),
            l3=F.Convolution2D(64, 64, ksize=3, stride=1, nobias=False, wscale=np.sqrt(2)),
            l4=F.Linear(3136, 512),#, wscale=np.sqrt(2)),
            q_value=F.Linear(512, n_act,
                             initialW=0.0*np.random.randn(n_act, 512).astype(np.float32))
        )
Example #10
Source File: dqn_agent_nips.py    From DQN-chainer with MIT License 5 votes vote down vote up 
def __init__(self, enable_controller=[0, 3, 4]):
        self.num_of_actions = len(enable_controller)
        self.enable_controller = enable_controller  # Default setting : "Pong"

        print "Initializing DQN..."
#	Initialization for Chainer 1.1.0 or older.
#        print "CUDA init"
#        cuda.init()

        print "Model Building"
        self.model = FunctionSet(
            l1=F.Convolution2D(4, 16, ksize=8, stride=4, wscale=np.sqrt(2)),
            l2=F.Convolution2D(16, 32, ksize=4, stride=2, wscale=np.sqrt(2)),
            l3=F.Linear(2592, 256),
            q_value=F.Linear(256, self.num_of_actions,
                             initialW=np.zeros((self.num_of_actions, 256),
                                               dtype=np.float32))
        ).to_gpu()

        print "Initizlizing Optimizer"
        self.optimizer = optimizers.RMSpropGraves(lr=0.0002, alpha=0.3, momentum=0.2)
        self.optimizer.setup(self.model.collect_parameters())

        # History Data :  D=[s, a, r, s_dash, end_episode_flag]
        self.D = [np.zeros((self.data_size, 4, 84, 84), dtype=np.uint8),
                  np.zeros(self.data_size, dtype=np.uint8),
                  np.zeros((self.data_size, 1), dtype=np.int8),
                  np.zeros((self.data_size, 4, 84, 84), dtype=np.uint8),
                  np.zeros((self.data_size, 1), dtype=np.bool)]
Example #11
Source File: dqn_agent_nature.py    From DQN-chainer with MIT License 5 votes vote down vote up 
def __init__(self, enable_controller=[0, 3, 4]):
        self.num_of_actions = len(enable_controller)
        self.enable_controller = enable_controller  # Default setting : "Pong"

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

        print "Model Building"
        self.model = FunctionSet(
            l1=F.Convolution2D(4, 32, ksize=8, stride=4, nobias=False, wscale=np.sqrt(2)),
            l2=F.Convolution2D(32, 64, ksize=4, stride=2, nobias=False, wscale=np.sqrt(2)),
            l3=F.Convolution2D(64, 64, ksize=3, stride=1, nobias=False, wscale=np.sqrt(2)),
            l4=F.Linear(3136, 512, wscale=np.sqrt(2)),
            q_value=F.Linear(512, self.num_of_actions,
                             initialW=np.zeros((self.num_of_actions, 512),
                                               dtype=np.float32))
        ).to_gpu()

        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.collect_parameters())

        # History Data :  D=[s, a, r, s_dash, end_episode_flag]
        self.D = [np.zeros((self.data_size, 4, 84, 84), dtype=np.uint8),
                  np.zeros(self.data_size, dtype=np.uint8),
                  np.zeros((self.data_size, 1), dtype=np.int8),
                  np.zeros((self.data_size, 4, 84, 84), dtype=np.uint8),
                  np.zeros((self.data_size, 1), dtype=np.bool)]
Example #12
Source File: q_net.py    From deel with MIT License 5 votes vote down vote up 
def __init__(self, use_gpu, enable_controller, dim):
		self.use_gpu = use_gpu
		self.num_of_actions = len(enable_controller)
		self.enable_controller = enable_controller
		self.dim = dim

		print("Initializing Q-Network...")

		hidden_dim = 256
		self.model = FunctionSet(
			l4=F.Linear(self.dim*self.hist_size, hidden_dim, wscale=np.sqrt(2)),
			q_value=F.Linear(hidden_dim, self.num_of_actions,
							 initialW=np.zeros((self.num_of_actions, hidden_dim),
											   dtype=np.float32))
		)
		if self.use_gpu >= 0:
			self.model.to_gpu()

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

		self.optimizer = optimizers.RMSpropGraves(lr=0.00025, alpha=0.95, momentum=0.95, eps=0.0001)
		self.optimizer.setup(self.model.collect_parameters())

		# History Data :  D=[s, a, r, s_dash, end_episode_flag]
		self.d = [np.zeros((self.data_size, self.hist_size, self.dim), dtype=np.uint8),
				  np.zeros(self.data_size, dtype=np.uint8),
				  np.zeros((self.data_size, 1), dtype=np.int8),
				  np.zeros((self.data_size, self.hist_size, self.dim), dtype=np.uint8),
				  np.zeros((self.data_size, 1), dtype=np.bool)]