tutorial_llm_gp_sr.yml

fitness_function:
        name: "alfa_ec_llm.problem_environments.symbolic_regression.SymbolicRegressionGPPlusSomeLLMConstrainedFewShot"

        # Name of file that has data(inputs and outputs) stored in it.
        fitness_cases_file: tests/data/fitness_cases.csv

        # Ratio of fitness cases used for training individual solutions
        test_train_split: 0.7

        # Number of evaluation exemplars to draw without replacement.
        # Mainly due to prompt limitations
        n_exemplars: 10

        # Number of samples provided to prompt
        n_shots: 2

        # All inputs and constants are automatically also set to terminals as
        # they have 0 arity. Variables are read from the fitness case file
        arities:
            +: 2
            '*': 2
            -: 2

        # Specify the constants you would like to employ in your evaluation
        # the program will use 0 to the number you specified
        constants:
            - 0
            - 1

# Specify the number of individuals(population) make sure the
# population size is GREATER than the number of inputs
population_size: 2

# Specify the max depth of the tree
max_depth: 5

# Specify the number of best individuals that are preserved between generations
elite_size: 1

# Specify the number of generations you would like to generate
generations: 2

# Specify the number of individual solutions that are compared when
# determining which solutions are inserted into next generation of the
# search loop
tournament_size: 2

# Random seed. Enables a user to replicate experiment
seed: 1

# Probability of individual solutions to be varied by the crossover operator
crossover_probability: 0.8

# Probability of individual solutions to be varied by the mutation operator
mutation_probability: 0.2