FlexMF WARP

This page analyzes the hyperparameter tuning results for the FlexMF scorer in implicit-feedback mode with WARP loss.

Parameter Search Space

Parameter	Type	Distribution	Values	Selected
embedding_size_exp	Integer	Uniform	3 ≤ \(x\) ≤ 10	6
regularization	Float	LogUniform	0.0001 ≤ \(x\) ≤ 10	0.211
learning_rate	Float	LogUniform	0.001 ≤ \(x\) ≤ 0.1	0.0781
reg_method	Categorical	Uniform	L2, AdamW	AdamW
item_bias	Categorical	Uniform	True, False	True

Final Result

Searching selected the following configuration:

{
    'embedding_size_exp': 6,
    'regularization': 0.21080704579971982,
    'learning_rate': 0.0780646991058362,
    'reg_method': 'AdamW',
    'item_bias': True,
    'epochs': 13
}

With these metrics:

{
    'RBP': 0.1317286837759703,
    'DCG': 1.4334711164614036,
    'NDCG': 0.40247882304526417,
    'RecipRank': 0.4149979501124277,
    'Hit10': 0.7566137566137566,
    'max_epochs': 50,
    'epoch_train_s': 0.11083783418871462,
    'epoch_measure_s': 0.23467652103863657,
    'done': False,
    'training_iteration': 13,
    'trial_id': '72878258',
    'date': '2025-09-30_18-03-55',
    'timestamp': 1759269835,
    'time_this_iter_s': 0.34922122955322266,
    'time_total_s': 4.267271280288696,
    'pid': 478906,
    'hostname': 'CCI-ws21',
    'node_ip': '10.248.127.152',
    'config': {
        'embedding_size_exp': 6,
        'regularization': 0.21080704579971982,
        'learning_rate': 0.0780646991058362,
        'reg_method': 'AdamW',
        'item_bias': True,
        'epochs': 13
    },
    'time_since_restore': 4.267271280288696,
    'iterations_since_restore': 13
}

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?