The API is new and often updated. Make heavy use of python's help function. So, for instance

from gpcam.gp_optimizer import fvGPOptimizer

gpo = fvGPOptimizer(...)


fvGPOptimizer inherits all functionalities from the gpOptimizer class, so have a look there for most of the methods methods. Some methods have been overwritten and explained below.

class fvGPOptimizer(input_space_dimensions, output_space_dimensions, output_number, input_space_bounds)


  • input_space_dimensions (int): the number of dimensions in the input space

  • output_space_dimensions (int): the number of dimensions in the output space

  • output_number (int): how many values will be output per data point in the input


  • index_set_bounds (2d np.array): bounds of the combined index set

Class Methods: This class inherits all methods from the GPOptimizer class. That is why there are only two changes to consider:

tell(x, y, variances = None, value_positions = None)

Same as in GPOptimizer with the addition of value_positions. Value positions is an np.array((len(data set),number of returns, dim of return))

init_fvgp(init_hyperparameters, compute_device = "cpu", gp_kernel_function = None, gp_mean_function = None,

sparse = False)

Have suggestions for the API or found a bug?

Write a brief description of what you saw or what you'd like to see.