| lm_file=5gram.lm.gz |
| |
| tm_file=dev.zh.grammar |
| tm_format=hiero |
| |
| glue_file=hiero.glue |
| glue_format=hiero |
| |
| #lm config |
| use_srilm=true |
| lm_ceiling_cost=100 |
| use_left_equivalent_state=false |
| use_right_equivalent_state=false |
| order=3 |
| |
| |
| #tm config |
| span_limit=10 |
| phrase_owner=pt |
| mono_owner=mono |
| begin_mono_owner=begin_mono |
| default_non_terminal=X |
| goalSymbol=S |
| |
| #pruning config |
| fuzz1=0.1 |
| fuzz2=0.1 |
| max_n_items=30 |
| relative_threshold=10.0 |
| max_n_rules=50 |
| rule_relative_threshold=10.0 |
| |
| #nbest config |
| use_unique_nbest=true |
| use_tree_nbest=false |
| add_combined_cost=true |
| top_n=300 |
| |
| |
| #remote lm server config, we should first prepare remote_symbol_tbl before starting any jobs |
| use_remote_lm_server=false |
| remote_symbol_tbl=./voc.remote.sym |
| num_remote_lm_servers=4 |
| f_remote_server_list=./remote.lm.server.list |
| remote_lm_server_port=9000 |
| |
| |
| #parallel deocoder: it cannot be used together with remote lm |
| num_parallel_decoders=1 |
| parallel_files_prefix=/tmp/ |
| |
| |
| ###### model weights |
| #lm order weight |
| lm 1.0 |
| |
| #phrasemodel owner column(0-indexed) weight |
| phrasemodel pt 0 0.5 |
| phrasemodel pt 1 0.5 |
| phrasemodel pt 2 0.5 |
| |
| #arityphrasepenalty owner start_arity end_arity weight |
| #arityphrasepenalty pt 0 0 1.0 |
| #arityphrasepenalty pt 1 1 -1.0 |
| #arityphrasepenalty pt 2 2 -2.0 |
| |
| #arityphrasepenalty glue 1 1 1.0 |
| #arityphrasepenalty glue 2 2 2.0 |
| |
| #phrasemodel mono 0 0.5 |
| |
| #wordpenalty weight |
| wordpenalty -1.0 |
| |
| #========= specify sparse feature def file here(this line should match the line in "params.txt.example") ========== |
| discriminative ./sparse_feat.example 1.0 |
| |
| #general |
| maxNumIter=10 |
| useSemiringV2=true |
| maxNumHGInQueue=40 |
| numThreads=20 |
| printFirstN=10 |
| |
| #option for first feature (e.g., baseline feature) |
| normalizeByFirstFeature=true |
| fixFirstFeature=false |
| |
| #loss-augmented pruning |
| lossAugmentedPrune=false |
| startLossScale=10 |
| lossDecreaseConstant=2 |
| |
| |
| #google linear corpus gain |
| useGoogleLinearCorpusGain=false |
| #googleBLEUWeights=-1.0;0.10277777076514476;0.07949965001350584;0.6993000659479868;0.09565585699195878 |
| googleBLEUWeights=-1.0;0.2941176470588235;0.42016806722689076;0.6002400960384154;0.8574858514834507 |
| |
| #annealing? |
| #0: no annealing at all; 1: quenching only; 2: cooling and then quenching |
| annealingMode=0 |
| |
| isScalingFactorTunable=false |
| useL2Regula=false |
| varianceForL2=1 |
| useModelDivergenceRegula=false |
| lambda=-1 |
| |
| #feature related |
| #dense features |
| useBaseline=false |
| baselineFeatureName=baseline_lzf |
| baselineFeatureWeight=1.0 |
| |
| useIndividualBaselines=true |
| baselineFeatIDsToTune=0;1 |
| |
| #sparse features |
| useSparseFeature=true |
| |
| useRuleIDName=false |
| |
| useTMFeat=false |
| useTMTargetFeat=false |
| |
| useMicroTMFeat=true |
| wordMapFile=/media/Data/JHU/Research/MT_discriminative_LM_training/joshua_expbleu/TEST/mr/wordMap |
| |
| useLMFeat=false |
| startNgramOrder=1 |
| endNgramOrder=2 |
