Skip to content

[160526]penn2malt wsj

Jiaqi Li edited this page Aug 3, 2016 · 1 revision

Machine: linearb

Branch: master

Command: spark-submit --driver-library-path $LD_LIBRARY_PATH --driver-memory 20g --executor-memory 20g --master 'local[*]' glm_parser.py -i 5 -s 8 -p ~/Daten/glm-parser-data/penn-wsj-deps/ --train='wsj_0[2-9][0-9][0-9].mrg.3.pa.gs.tab|wsj_1[0-9][0-9][0-9].mrg.3.pa.gs.tab|wsj_2[0-1][0-9][0-9].mrg.3.pa.gs.tab' --test='wsj_0[0-1][0-9][0-9].mrg.3.pa.gs.tab|wsj_22[0-9][0-9].mrg.3.pa.gs.tab|wsj_24[0-9][0-9].mrg.3.pa.gs.tab' -d 27-05-2016 -a --learner=average_perceptron --fgen=english_1st_fgen --parser=ceisner --config=config/penn2malt.config

Result: 05/27/2016 03:09:04 PM DEBUG: Initialize AveragePerceptronLearner ... 05/27/2016 03:09:07 PM DEBUG: Initialize ParallelPerceptronLearner ... 05/28/2016 01:46:27 AM DEBUG: Start evaluating ... 05/28/2016 01:58:02 AM INFO: Training time usage(seconds): 38243.774359 05/28/2016 01:58:02 AM INFO: Feature count: 8195037 05/28/2016 01:58:02 AM INFO: Unlabeled accuracy: 0.886755180959 (148136, 167054) 05/28/2016 01:58:02 AM INFO: Unlabeled attachment accuracy: 0.891284609284 (155096, 174014)

Clone this wiki locally