Python baselines.common.tf_util.numel() Examples

The following are 23 code examples of baselines.common.tf_util.numel(). 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 baselines.common.tf_util , or try the search function .
Example #1
Source File: util.py    From ICML2019-TREX with MIT License 5 votes vote down vote up 
def flatten_grads(var_list, grads):
    """Flattens a variables and their gradients.
    """
    return tf.concat([tf.reshape(grad, [U.numel(v)])
                      for (v, grad) in zip(var_list, grads)], 0)
Example #2
Source File: mpi_adam.py    From BackpropThroughTheVoidRL with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None else comm
Example #3
Source File: mpi_adam.py    From deeprl-baselines with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None else comm
Example #4
Source File: util.py    From Overcoming-exploration-from-demos with MIT License 5 votes vote down vote up 
def flatten_grads(var_list, grads):
    """Flattens a variables and their gradients.
    """
    return tf.concat([tf.reshape(grad, [U.numel(v)])
                      for (v, grad) in zip(var_list, grads)], 0)
Example #5
Source File: mpi_adam.py    From baselines with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None and MPI is not None else comm
Example #6
Source File: util.py    From baselines with MIT License 5 votes vote down vote up 
def flatten_grads(var_list, grads):
    """Flattens a variables and their gradients.
    """
    return tf.concat([tf.reshape(grad, [U.numel(v)])
                      for (v, grad) in zip(var_list, grads)], 0)
Example #7
Source File: mpi_adam.py    From self-imitation-learning with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None else comm
Example #8
Source File: util.py    From self-imitation-learning with MIT License 5 votes vote down vote up 
def flatten_grads(var_list, grads):
    """Flattens a variables and their gradients.
    """
    return tf.concat([tf.reshape(grad, [U.numel(v)])
                      for (v, grad) in zip(var_list, grads)], 0)
Example #9
Source File: mpi_adam.py    From sonic_contest with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None else comm
Example #10
Source File: util.py    From sonic_contest with MIT License 5 votes vote down vote up 
def flatten_grads(var_list, grads):
    """Flattens a variables and their gradients.
    """
    return tf.concat([tf.reshape(grad, [U.numel(v)])
                      for (v, grad) in zip(var_list, grads)], 0)
Example #11
Source File: mpi_adam.py    From MOREL with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None else comm
Example #12
Source File: mpi_adam.py    From ICML2019-TREX with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None and MPI is not None else comm
Example #13
Source File: util.py    From lirpg with MIT License 5 votes vote down vote up 
def flatten_grads(var_list, grads):
    """Flattens a variables and their gradients.
    """
    return tf.concat([tf.reshape(grad, [U.numel(v)])
                      for (v, grad) in zip(var_list, grads)], 0)
Example #14
Source File: mpi_adam.py    From ICML2019-TREX with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None and MPI is not None else comm
Example #15
Source File: util.py    From ICML2019-TREX with MIT License 5 votes vote down vote up 
def flatten_grads(var_list, grads):
    """Flattens a variables and their gradients.
    """
    return tf.concat([tf.reshape(grad, [U.numel(v)])
                      for (v, grad) in zip(var_list, grads)], 0)
Example #16
Source File: mpi_adam.py    From DRL_DeliveryDuel with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None else comm
Example #17
Source File: util.py    From DRL_DeliveryDuel with MIT License 5 votes vote down vote up 
def flatten_grads(var_list, grads):
    """Flattens a variables and their gradients.
    """
    return tf.concat([tf.reshape(grad, [U.numel(v)])
                      for (v, grad) in zip(var_list, grads)], 0)
Example #18
Source File: mpi_adam.py    From learning2run with MIT License 5 votes vote down vote up 
def __init__(self, var_list, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None else comm
Example #19
Source File: mpi_adam.py    From rl_graph_generation with BSD 3-Clause "New" or "Revised" License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None else comm
Example #20
Source File: util.py    From rl_graph_generation with BSD 3-Clause "New" or "Revised" License 5 votes vote down vote up 
def flatten_grads(var_list, grads):
    """Flattens a variables and their gradients.
    """
    return tf.concat([tf.reshape(grad, [U.numel(v)])
                      for (v, grad) in zip(var_list, grads)], 0)
Example #21
Source File: mpi_adam.py    From HardRLWithYoutube with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None else comm
Example #22
Source File: util.py    From HardRLWithYoutube with MIT License 5 votes vote down vote up 
def flatten_grads(var_list, grads):
    """Flattens a variables and their gradients.
    """
    return tf.concat([tf.reshape(grad, [U.numel(v)])
                      for (v, grad) in zip(var_list, grads)], 0)
Example #23
Source File: mpi_adam.py    From lirpg with MIT License 5 votes vote down vote up 
def __init__(self, var_list, *, beta1=0.9, beta2=0.999, epsilon=1e-08, scale_grad_by_procs=True, comm=None):
        self.var_list = var_list
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.scale_grad_by_procs = scale_grad_by_procs
        size = sum(U.numel(v) for v in var_list)
        self.m = np.zeros(size, 'float32')
        self.v = np.zeros(size, 'float32')
        self.t = 0
        self.setfromflat = U.SetFromFlat(var_list)
        self.getflat = U.GetFlat(var_list)
        self.comm = MPI.COMM_WORLD if comm is None else comm