Initial guess

[vthorey]

Initial guess

[vthorey]

Hi, You can use

optimization_problem = OptimizationProblem.from_list([
    {
        "name": "param1",
        "category": "category",
        "search_space": {"values": ["a", "b"]}
    },
    {
        "name": "param1",
        "category": "normal",
        "search_space": {"values": [1, 2]}
    }
])

optimization_problem.add_observations_from_list(
    [
        {
            "loss": 0.5,
            "parameters": {"param1": "a", "param2": 1.5}
        },
        {
            "loss": 0.7,
            "parameters": {"param1": "a", "param2": 1.8}
        },
        {
            "loss": 0.1,
            "parameters": {"param1": "b", "param2": 1.1}
        },
]

[c0d3rman]

Ah, I think I see my issue. I was trying to use minimize, but it seems like that function only accepts a list, and throws an error (UnboundLocalError: local variable 'optimization_problem' referenced before assignment) if it's passed in anything else. If I implement minimize myself it works fine. All it takes is copying the code from minimize and replacing all occurrences of optimization_problem_parameters with optimization_problem. It seems like this was the original intention for minimize?

[vthorey]

You are right, It is clarified now thanks!