cdec cfg 'cdec.ini' Loading the LM will be faster if you build a binary file. Reading ../example/nc-wmt11.en.srilm.gz ----5---10---15---20---25---30---35---40---45---50---55---60---65---70---75---80---85---90---95--100 **************************************************************************************************** Seeding random number sequence to 2767202922 dtrain Parameters: k 100 N 4 T 1 scorer 'stupid_bleu' sample from 'kbest' filter 'uniq' learning rate 0.0001 gamma 0 loss margin 1 pairs 'XYX' hi lo 0.1 pair threshold 0 select weights 'last' l1 reg 0 'none' max pairs 4294967295 cdec cfg 'cdec.ini' input 'work/shard.0.0.in' refs 'work/shard.0.0.refs' output 'work/weights.0.1' weights in 'work/weights.0' (a dot represents 10 inputs) Iteration #1 of 1. 5 WEIGHTS Glue = -0.2699 WordPenalty = +0.080605 LanguageModel = -0.026572 LanguageModel_OOV = -0.30025 PhraseModel_0 = -0.32076 PhraseModel_1 = +0.67451 PhraseModel_2 = +0.92 PhraseModel_3 = -0.36402 PhraseModel_4 = -0.592 PhraseModel_5 = -0.0269 PhraseModel_6 = -0.28755 PassThrough = -0.33285 --- 1best avg score: 0.26638 (+0.26638) 1best avg model score: 53.197 (+53.197) avg # pairs: 2028.6 avg # rank err: 998.2 avg # margin viol: 918.8 non0 feature count: 12 avg list sz: 100 avg f count: 10.496 (time 0.32 min, 3.8 s/S) Writing weights file to 'work/weights.0.1' ... done --- Best iteration: 1 [SCORE 'stupid_bleu'=0.26638]. This took 0.31667 min.