FlexMF BPR

This page analyzes the hyperparameter tuning results for the FlexMF scorer in implicit-feedback mode with pairwise loss (Bayesian Personalized Ranking).

Parameter Search Space

/home/mde48/lenskit/lenskit-codex/.venv/lib/python3.12/site-packages/ray/tune/search/sample.py:700: RayDeprecationWarning: The `base` argument is deprecated. Please remove it as it is not actually needed in this method.

Parameter	Type	Distribution	Values	Selected
embedding_size	Integer	LogUniform	4 ≤ \(x\) ≤ 512	8
regularization	Float	LogUniform	0.0001 ≤ \(x\) ≤ 10	0.0463
learning_rate	Float	LogUniform	0.001 ≤ \(x\) ≤ 0.1	0.00508
reg_method	Categorical	Uniform	L2, AdamW	AdamW
negative_count	Integer	Uniform	1 ≤ \(x\) ≤ 5	4
item_bias	Categorical	Uniform	True, False	True

Final Result

Searching selected the following configuration:

{
    'embedding_size': 8,
    'regularization': 0.04628329414535743,
    'learning_rate': 0.005081316811240104,
    'reg_method': 'AdamW',
    'negative_count': 4,
    'item_bias': True,
    'epochs': 12
}

With these metrics:

{
    'RBP': 0.17365074017826568,
    'LogRBP': 2.0435307324426404,
    'NDCG': 0.41911113610176876,
    'RecipRank': 0.3369176769793785,
    'TrainTask': '577f0b6d-26ff-40e6-964f-1afddac35523',
    'TrainTime': None,
    'TrainCPU': None,
    'max_epochs': 50,
    'done': True,
    'training_iteration': 12,
    'trial_id': '6a9e4126',
    'date': '2025-05-07_06-26-18',
    'timestamp': 1746613578,
    'time_this_iter_s': 48.02631378173828,
    'time_total_s': 713.4712646007538,
    'pid': 210716,
    'hostname': 'CCI-ws21',
    'node_ip': '10.248.127.152',
    'config': {
        'embedding_size': 8,
        'regularization': 0.04628329414535743,
        'learning_rate': 0.005081316811240104,
        'reg_method': 'AdamW',
        'negative_count': 4,
        'item_bias': True,
        'epochs': 12
    },
    'time_since_restore': 48.02631378173828,
    'iterations_since_restore': 1
}