How to correctly predict using best model in talos

Question

I managed to use talos scan to generate model grid but I can't use the evaluate function

SCANNING

    t = talos.Scan(x=x,
                   y=y,
                   params=p,
                   shuffle=False,
                   x_val=x_val_multi,
                   y_val=y_val_multi,
                   model=sonos,
                   experiment_name='exp1',
                   disable_progress_bar=False,
                   print_params=False)

EVALUATION

e = Evaluate(t)
e.evaluate(x_val, 
                y_val,  
                model_id=None,
                folds=10,
                shuffle=True,
                metric='val_acc',
                asc=False,
                print_out=False)

but it keeps throwing

TypeError: evaluate() missing 1 required positional argument: 'task'

Answer 1

Must specify task option with string argument. Options are binary, multi_class, multi_label, and continuous

e.g.:

e = Evaluate(t)
e.evaluate(x_val, 
                y_val,  
                model_id=None,
                folds=10,
                shuffle=True,
                metric='val_acc',
                asc=False,
                print_out=False,
                task = 'binary')