rime.Experiment

class rime.Experiment(D, V=None, *V_extra, default_k_items_per_user=None, default_c_users_per_item=None, mult=[], models_to_run=None, model_hyps={}, device='cpu', dual=False, online=False, tie_break=0, cache=None, results=None, **mtch_kw)

Bases: object

Produce item_rec / user_rec metrics; then sweeps through multipliers for relevance-diversity curve, interpreting mult<1 as item min-exposure and mult>=1 as user max-limit

__init__(D, V=None, *V_extra, default_k_items_per_user=None, default_c_users_per_item=None, mult=[], models_to_run=None, model_hyps={}, device='cpu', dual=False, online=False, tie_break=0, cache=None, results=None, **mtch_kw)

Methods

__init__(D[, V, default_k_items_per_user, ...])
metrics_update(name, S[, T])
run([models_to_run])	models_to_exclude is ignored if models_to_run is explicitly provided
update_cache(other)

Attributes

registered

run(models_to_run=None)

models_to_exclude is ignored if models_to_run is explicitly provided