cbe_adf_cover_e_dec_predict.stderr

only testing
predictions = cbe_adf_cover_e_dec.predict
warning: currently, mtr is only used for the first policy in cover, other policies use dr
Num weight bits = 18
learning rate = 0.5
initial_t = 0
power_t = 0.5
using no cache
Reading datafile = train-sets/cb_test_medium.ldf
num sources = 1
Enabled reductions: gd, scorer, csoaa_ldf, cb_adf, cb_explore_adf_cover, shared_feature_merger
average  since         example        example  current  current  current
loss     last          counter         weight    label  predict features
0.365148 0.365148            1            1.0  0:1:0.5        2:0.634852...        9
0.182574 0.000000            2            2.0  1:0:0.5        1:0.666667...        6
0.243432 0.365148            4            4.0  0:1:0.5        2:0.634852...        9
0.182574 0.121716            8            8.0  1:0:0.5        1:0.666667...        6

finished run
number of examples = 15
weighted example sum = 15.000000
weighted label sum = 0.000000
average loss = 0.182574
total feature number = 115