tf.uniform_unit_scaling_initializer

class tf.uniform_unit_scaling_initializer

See the guide: Variables > Sharing Variables

Initializer that generates tensors without scaling variance.

When initializing a deep network, it is in principle advantageous to keep the scale of the input variance constant, so it does not explode or diminish by reaching the final layer. If the input is x and the operation x * W, and we want to initialize W uniformly at random, we need to pick W from

[-sqrt(3) / sqrt(dim), sqrt(3) / sqrt(dim)]

to keep the scale intact, where dim = W.shape[0] (the size of the input). A similar calculation for convolutional networks gives an analogous result with dim equal to the product of the first 3 dimensions. When nonlinearities are present, we need to multiply this by a constant factor. See Sussillo et al., 2014 (pdf) for deeper motivation, experiments and the calculation of constants. In section 2.3 there, the constants were numerically computed: for a linear layer it's 1.0, relu: ~1.43, tanh: ~1.15.

Args:

  • factor: Float. A multiplicative factor by which the values will be scaled.
  • seed: A Python integer. Used to create random seeds. See tf.set_random_seed for behavior.
  • dtype: The data type. Only floating point types are supported.

Methods

__init__(factor=1.0, seed=None, dtype=tf.float32)

Defined in tensorflow/python/ops/init_ops.py.