Skip to content

2016_10_1_pos_results

Jump to bottom Edit New page
Vivian Kou edited this page Oct 4, 2016 · 1 revision

Part of speech tagging:

  1. train on 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 with 8 shards and 10 iterations

  2. test on 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

  3. results: 2016-09-30 13:14:48,408 TAGGER [INFO]: Starting evaluation process 2016-09-30 13:14:50,668 EVALUATOR [INFO]: Totel number of sentences: 6960 2016-09-30 14:22:09,084 EVALUATOR [INFO]: Feature count: 500856 2016-09-30 14:22:09,084 EVALUATOR [INFO]: Unlabelled accuracy: 0.970736127842 (175678, 180974) 2016-09-30 14:22:09,084 EVALUATOR [INFO]: Unlabelled attachment accuracy: 0.971819894218 (182638, 187934) 2016-09-30 14:22:09,103 TAGGER [INFO]: Total evaluation Time(seconds): 4040.694758

Add a custom footer
Clone this wiki locally