Generators

Gaussian Generator

layer eddl::GaussGenerator(float mean, float stdev, vector<int> size)

Generates a gaussian noise output (typically used for GANs) with the specified mean and standard deviation.

Parameters
  • mean – Mean of the gaussian distribution

  • stdev – Standard deviation of the gaussian distribution

  • size – Shape of the output tensor of the layer

Returns

A layer that generates tensors with the specified gaussian distribution.

Example:

layer gin = GaussGenerator(0.0, 1, {3, 32, 32});

Uniform Generator

layer eddl::UniformGenerator(float low, float high, vector<int> size)

Generates a uniform noise output (typically used for GANs) with the specified lower and upper bound values.

Parameters
  • low – Lower bound of the uniform distribution

  • high – Upper bound of the uniform distribution

  • size – Shape of the output tensor of the layer

Returns

A layer that generates tensors with the specified uniform distribution.

layer gin = UniformGenerator(3.0, 10.0, {3, 32, 32});