Skip to content

Latest commit

 

History

History
11 lines (10 loc) · 291 Bytes

README.md

File metadata and controls

11 lines (10 loc) · 291 Bytes

BERT_chinese_LM_processing

  • legal text similarity
  • finetune and extract feature from regulation data. (zh_TW TF2.0)

Methodology

  • ensemble (domain specific)
    • diffrent tokenizer
    • diffrent embedding
    • different pretrained
  • 0X.For PoC THUCNews
  • 00X. For old data sinopac