cb_zero_feature_vals_dsjson.stderr

predictions = cb_zero_feature_vals_dsjson.predict
Num weight bits = 18
learning rate = 0.5
initial_t = 0
power_t = 0.5
using no cache
Reading datafile = train-sets/cb_features_w_zero_vals.dsjson
WARNING: Old string feature value behavior is deprecated in JSON/DSJSON and will be removed in a future version. Use `--chain_hash` to use new behavior and silence this warning.
num sources = 1
Enabled reductions: gd, scorer, csoaa_ldf, cb_adf, cb_explore_adf_greedy, shared_feature_merger
average  since         example        example  current  current  current
loss     last          counter         weight    label  predict features
-1.000000 -1.000000            1            1.0 0:-1:0.5        0:0.5...        2
-1.000000 -1.000000            2            2.0 0:-1:0.5        0:0.5...        2

finished run
number of examples = 2
weighted example sum = 2.000000
weighted label sum = 0.000000
average loss = -1.000000
total feature number = 6