FlexMF Explicit

This page analyzes the hyperparameter tuning results for the FlexMF scorer in explicit-feedback mode (a biased matrix factorization model trained with PyTorch).

Parameter Search Space

Parameter	Type	Distribution	Values	Selected
embedding_size_exp	Integer	Uniform	3 ≤ \(x\) ≤ 10	5
regularization	Float	LogUniform	0.0001 ≤ \(x\) ≤ 10	0.153
learning_rate	Float	LogUniform	0.001 ≤ \(x\) ≤ 0.1	0.00166
reg_method	Categorical	Uniform	L2, AdamW	AdamW

Final Result

Searching selected the following configuration:

{
    'embedding_size_exp': 5,
    'regularization': 0.15268775865258827,
    'learning_rate': 0.0016555960135169957,
    'reg_method': 'AdamW',
    'epochs': 6
}

With these metrics:

{
    'RBP': 0.16262047445215053,
    'DCG': 10.479418851290816,
    'NDCG': 0.3969117219917976,
    'RecipRank': 0.33687097501755486,
    'Hit10': 0.5717171717171717,
    'RMSE': 0.7818160653114319,
    'max_epochs': 50,
    'epoch_train_s': 1.3008685240056366,
    'epoch_measure_s': 13.93932611704804,
    'done': True,
    'training_iteration': 6,
    'trial_id': '46cb36ae',
    'date': '2025-10-01_11-13-15',
    'timestamp': 1759331595,
    'time_this_iter_s': 15.24461555480957,
    'time_total_s': 88.9127516746521,
    'pid': 1114212,
    'hostname': 'CCI-ws21',
    'node_ip': '10.248.127.152',
    'config': {
        'embedding_size_exp': 5,
        'regularization': 0.15268775865258827,
        'learning_rate': 0.0016555960135169957,
        'reg_method': 'AdamW',
        'epochs': 6
    },
    'time_since_restore': 88.9127516746521,
    'iterations_since_restore': 6
}

Parameter Analysis

Embedding Size

The embedding size is the hyperparameter that most affects the model’s fundamental logic, so let’s look at performance as a fufnction of it:

Iteration Completion

How many iterations, on average, did we complete?

How did the metric progress in the best result?

How did the metric progress in the longest results?