ymcui/Chinese-BERT-wwm

Issues

• 您好，我想请教一个代码上的细节

  #243 opened 3 months ago by 4vicii
  0

• 有点晕：『开源版本不包含MLM任务的权重』,这句话是什么意思？

  #226 opened 2 years ago by liangcaihua
  3

• RoBERTa-wwm-ext-large应用到全新领域不收敛

  #227 opened 2 years ago by JerryYao80
  4

• 从BERT换成此模型后跑不通，应该怎么解决，显示缺少参数

  #242 opened a year ago by ottt08
  1

• 请问每个模型推理时需要多大的显存去加载？

  #240 opened a year ago by JohnHerry
  1

• 请问pytorch版本的chinese-roberta-wwm-ext-large在哪里下载稳定而且快？huggingface上太慢了还总断。

  #241 opened a year ago by JohnHerry
  1

• Is onnx model available

  #239 opened a year ago by panxin801
  2

• Is there any sharing about phoneme-BERT pretrained?

  #238 opened a year ago by JohnHerry
  2

• 我用bert官方的分类代码去跑，用bert-wwm作为预训练模型，报错TypeError: __init__() takes 1 positional argument but 3 were given

  #209 opened 3 years ago by dolphin-Jia
  3

• NER问题

  #237 opened a year ago by wanghao-007
  2

• 请问可否使用wwm系列的模型做词级别的fill-mask预测？

  #236 opened a year ago by LiShaoyu5
  2

• 预训练数据

  #235 opened a year ago by rattlesnakey
  2

• 如何抽取特定layer的词向量？

  #234 opened a year ago by Black-Rhen
  2

• 继续预训练

  #233 opened a year ago by yyggano
  2

• 关于不能用RobertaModel引用

  #206 opened 3 years ago by Jackthebighead
  8

• RoBERTa-wwm-ext-large ft的时候loss飞了

  #232 opened 2 years ago by fword
  2

• How WWM stratege works in code

  #230 opened 2 years ago by DemonDamon
  2

• 链接失效求助

  #231 opened 2 years ago by Alternate-D
  2

• CJRC有预训练的模型么？

  #229 opened 2 years ago by ucas010
  1

• How to download chinese-roberta-wwm-ext.pt ?

  #228 opened 2 years ago by 2000yang
  1

• 预训练分词是会用到中文vocab后面带“##”的token吗？如果是，整词掩码预训练出来的语言模型，用在下游任务中，下游任务可以不分词吗？

  #224 opened 2 years ago by dr-GitHub-account
  4

• 请问一下训练用了几块tpu呀

  #223 opened 2 years ago by guohan950106
  2

• 预训练阶段的loss问题

  #222 opened 2 years ago by rtygbwwwerr
  2

• 请问有办法对现有中文整词掩码预训练模型做 adaptive fine-tuning 吗？

  #220 opened 2 years ago by dr-GitHub-account
  4

• wwm mask 细节

  #219 opened 2 years ago by zhengjiawei001
  2

• 关于词表没有中文双引号的问题

  #221 opened 2 years ago by Dicer-Zz
  3

• EXT数据集的量有多少

  #218 opened 2 years ago by xueyuan1990
  1

• “注意：开源版本不包含MLM任务的权重；如需做MLM任务，请进行二次预训练。”

  #216 opened 3 years ago by lxxXuan
  3

• 词汇级词向量

  #215 opened 3 years ago by wangcccchen
  1

• pad_token_id错误

  #214 opened 3 years ago by CaoYiwei
  5

• 请问BERT-wwm-ext有使用nsp任务吗？

  #213 opened 3 years ago by lionsterbenn
  2

• 计算两句子的相似度

  #205 opened 3 years ago by yfq512
  2

• 请问RoBERTa-wwm-ext有没有随机初始化MLM部分的权重

  #212 opened 3 years ago by dolphin-Jia
  2

• Confusion with the config.json in RoBerta-based Models

  #210 opened 3 years ago by qhd1996
  2

• 讯飞的下载链接提示外链已经过期了

  #211 opened 3 years ago by natureLanguageQing
  4

• The way to get Chinese Word Embedding.

  #208 opened 3 years ago by qhd1996
  4

• 有没有Bert-wwm-ext-large模型?

  #207 opened 3 years ago by zhaowei-wang-nlp
  2

• rbt3将tf模型转torch报错

  #204 opened 3 years ago by renjunxiang
  2

• model sizing of ”BERT-wwm-ext, Chinese” and ”BERT-wwm, Chinese”

  #203 opened 3 years ago by elulue
  2

• 求问不同模型的schedule细节

  #202 opened 3 years ago by NinedayWang
  2

• 求教 roberta既然在训练的时候未做NSP，那pooler_output 还有意义吗

  #201 opened 3 years ago by rmbone
  2

• wwm模型加载的时候tokenizer出来都是一个个的字，这样对吗？

  #200 opened 3 years ago by rmbone
  2

• 求助：谁有pytorch预训练模型的国内下载地址吗？我从Google云下不下来，谢谢！

  #199 opened 3 years ago by DaihuaWei
  3

• roberta中cls token的计算方法

  #198 opened 3 years ago by wikithink
  3

• 求教：为什么词语要对应多个mask？

  #197 opened 3 years ago by fangwc
  3

• > @beamind 请问复现结果了吗

  #195 opened 3 years ago by houjin0803
  2

• 序列标注任务警告信息

  #194 opened 3 years ago by WillionLei
  2

• 这个模型可以不做fine-tuning直接用吗？

  #193 opened 3 years ago by xiaohengheng
  2

• 如何获得词向量？

  #192 opened 3 years ago by c9412600
  2

• 实验结果复现问题

  #196 opened 3 years ago by houjin0803
  2