sequence_tagging_for_ner训练过程F1为什么一直是0?
Created by: utopiar
[INFO 2017-11-01 18:49:00,898 train.py:83] Pass 0, Batch 0, Cost 15.435392, {'ner_chunk.precision': 0.02575107291340828, 'ner_chunk.F1-score': 0.03539822995662689, 'ner_chunk.recall': 0.056603774428367615, 'error': 0.8953229188919067}
[INFO 2017-11-01 18:49:05,185 train.py:87]
Test with Pass 0, Batch 0, {'ner_chunk.precision': 0.019510731101036072, 'ner_chunk.F1-score': 0.029499361291527748, 'ner_chunk.recall': 0.06044403091073036, 'error': 0.9291946291923523}
[INFO 2017-11-01 18:49:05,246 train.py:83] Pass 0, Batch 1, Cost 11.822507, {'ner_chunk.precision': 0.0, 'ner_chunk.F1-score': 0.0, 'ner_chunk.recall': 0.0, 'error': 0.24517905712127686}
[INFO 2017-11-01 18:49:09,545 train.py:87]
Test with Pass 0, Batch 1, {'ner_chunk.precision': 0.0, 'ner_chunk.F1-score': 0.0, 'ner_chunk.recall': 0.0, 'error': 0.16679592430591583}
[INFO 2017-11-01 18:49:09,601 train.py:83] Pass 0, Batch 2, Cost 10.640020, {'ner_chunk.precision': 0.0, 'ner_chunk.F1-score': 0.0, 'ner_chunk.recall': 0.0, 'error': 0.29411765933036804}
[INFO 2017-11-01 18:49:13,849 train.py:87]
Test with Pass 0, Batch 2, {'ner_chunk.precision': 0.0, 'ner_chunk.F1-score': 0.0, 'ner_chunk.recall': 0.0, 'error': 0.16679592430591583}
除了第一个Batch的F1不是0,以后一直是0,请教这是什么问题呢?