blob: 2fa77b6c299284207d69ca559b639271ae5e511d (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
|
creating quadratic features for pairs: ui
final_regressor = movielens.reg
Num weight bits = 16
learning rate = 0.025
initial_t = 1
power_t = 0
decay_learning_rate = 0.97
rank = 10
using l2 regularization = 0.001
creating cache_file = movielens.cache
Reading datafile = train-sets/ml100k_small_train
num sources = 1
average since example example current current current
loss last counter weight label predict features
8.699942 8.699942 3 3.0 4.0000 0.2004 23
11.041393 13.382844 6 6.0 4.0000 0.3577 23
8.934545 6.406327 11 11.0 1.0000 0.6111 23
7.357914 5.781282 22 22.0 2.0000 1.2071 23
6.031830 4.705747 44 44.0 4.0000 1.7897 23
4.832065 3.604398 87 87.0 3.0000 2.9281 23
3.078799 1.325534 174 174.0 4.0000 3.4095 23
2.182642 1.286485 348 348.0 5.0000 3.7808 23
finished run
number of examples = 474
weighted example sum = 474
weighted label sum = 1666
average loss = 1.945
best constant = 3.52
total feature number = 10902
|
