xu-song/bert-as-language-model

Issues

• 请问模型微调过吗?

  #20 opened 10 months ago by RumRaisins
  2

• Probability of the last word is always too small

  #9 opened 5 years ago by AliceZhang2016
  2

• 推导结果不正常

  #19 opened a year ago by zhangyisheng0629
  2

• paper to cite

  #11 opened a year ago by JohannaOm
  1

• 求概率还是用AR model

  #18 opened 4 years ago by mlpod
  1

• 就是按顺序依次mask掉每一个词，然后预测该词的概率。

  #17 opened 4 years ago by duan348733684
  4

• online service的问题

  #16 opened 4 years ago by aaronliu7
  4

• TODO:

  #1 opened 4 years ago by xu-song
  1

• 原理是什么？

  #15 opened 4 years ago by aaronliu7
  2

• 每次执行时提示找不到模型

  #7 opened 4 years ago by barry1006
  10

• predict耗时太久，运行一条语句需要10s，请问怎么改进呢

  #8 opened 4 years ago by boy-be-ambitious
  3

• serving 的问题

  #12 opened 4 years ago by kylezhaoxc
  1

• 请教下作者xu-song， case里面中文句子打分，用的是分词后的模型，还是单字的模型呢。

  #13 opened 4 years ago by barry1006
  1

• 预训练模型有没有保存cls/predictions下的权重

  #14 opened 4 years ago by joytianya
  1

• 为什么要设置max_predictions_per_seq

  #10 opened 5 years ago by zhanghaoie
  2

• 作为language model对句子打分需要finetune吗

  #2 opened 6 years ago by HyacinthJingjing
  3

• 字prob和句子perplexity计算方法

  #5 opened 6 years ago by lxylxyoo
  2

• Words outside of vocab

  #4 opened 6 years ago by ziyaoh
  1

• Export bert-as-language-model as online service but failed.

  #3 opened 6 years ago by stevewyl
  3