brightmart roberta_zh issues

brightmart / roberta_zh

RoBERTa中文预训练模型: RoBERTa for Chinese

2.62k stars 409 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

updata:测试

#100 Cpeidong opened 3 months ago
1
RoBERTa_zh_Large_PyTorch的网盘链接失效了，能麻烦提供新的链接吗？

#99 zzzengzhe opened 9 months ago
3
加载的小问题求解答

#98 cooper12121 opened 1 year ago
0
下载问题和加载模型

#97 Lj4040 opened 1 year ago
0
333add

#96 luoshao23 closed 2 years ago
0
Loss curve

#95 wanglaiqi opened 2 years ago
0
tensorboard可视化模型输出结果 train的masked_lm_loss和masked_lm_accuracy是空的，eval的图只有一个点

#94 hhanyyan opened 2 years ago
0
利用roberta_zh的tokenizer来做中文NER任务时报错

#93 Honma-Rika opened 3 years ago
2
Update README.md

#92 Yang-Jianzhang closed 3 years ago
0
Huggingface

#91 archersama opened 3 years ago
0
请问下，怎么进行GPU训练？

#90 Tian14267 opened 3 years ago
0
pytorch用BERT的加载方式加载roberta模型，呢么创建token时special token 是按照bert的方式还是roberta的方式呢

#89 ludfeer opened 3 years ago
0
resource文件夹下的vocab和代码不对应

#88 bigheary closed 3 years ago
0
Unrelated parameters in the config

#87 zheyuye opened 3 years ago
0
预处理数据丢失问题

#86 puzzledTao opened 3 years ago
1
What are the pretrained-language-model that is obviously better than BERT and RoBERTa?

#85 guotong1988 opened 3 years ago
0
关于MLM中，中文全词掩盖的预测标签问题

#84 Rango94 opened 3 years ago
7
其中依赖的预训练模型是否和bert官方提供是一样的？

#83 charlesfufu opened 3 years ago
0
是否可以开放语料，供其他模型对比

#82 lamp-lyz opened 4 years ago
0
CMRC示例

#81 chuzhifeng closed 4 years ago
0
NaN probability sometimes when inference on GPU

#80 Jiayuforfreeo opened 4 years ago
0
在pytorch模型上做post train

#79 daniellibin opened 4 years ago
2
GPT vs BERT, under same computation and data resource, which one is better for downstream tasks like GLUE?

#78 guotong1988 opened 4 years ago
0
XLNet其实不能稳压RoBERTa吧？

#77 guotong1988 closed 4 years ago
1
做全词遮蔽的比例设置？

#76 surimj opened 4 years ago
1
关于中文编码

#75 Leputa opened 4 years ago
0
pretrain问题：ValueError: Please provide a TPU Name to connect to.

#74 xuehui0725 closed 4 years ago
2
pretrain 数据问题

#73 ruleGreen opened 4 years ago
3
预训练数据构造有误

#72 hy-struggle opened 4 years ago
0
Update create_pretraining_data.py

#71 guotong1988 opened 4 years ago
0
用自己的数据构建pretrain data 提示 KeyError: '##cry'

#70 ccoocode opened 4 years ago
1
希望能提供一个longformer的中文预训练模型

#69 xjx0524 opened 4 years ago
0
没有merge.txt和vocab.json

#68 lshowway opened 4 years ago
3
Checksum does not match ,请问是TensorFlow版本的问题吗，我的是1.15.0

#67 545314690 opened 4 years ago
0
动态mask逻辑的实现

#66 humdingers opened 4 years ago
1
部分参数无法重载

#65 WBY1993 opened 4 years ago
0
数据处理中re.findall('##[\u4E00-\u9FA5]')作用

#64 xiaojinglu opened 4 years ago
2
你好，谷歌云的roberta下载好像不太可用了？方便重新给出链接吗

#63 currywu123 opened 4 years ago
0
请问英文的roberta预训练模型哪里可以找得到？？？

#62 WenxiongLiao opened 4 years ago
1
如何在论文中引用您发布的模型？

#61 CCNUdhj opened 4 years ago
0
请问一下有如下的权重可以分享吗?

#60 rxc205 closed 4 years ago
1
tensorflow版本

#59 aflyhat opened 4 years ago
1
您好，我想问一下，在预训练的时候怎么做到多gpu并行？

#58 chenchengshuai opened 4 years ago
0
使用12层预训练roberta_zh_l12模型做类似于bert as service的句子embedding提取器报错

#57 guoraikkonen closed 4 years ago
2
您好，我用您训练好的模型在我的语料上进行微调没有任何问题。我做了两个实验，一个是用我自己的语料从头开始预训练，一个是用我的语料在您的模型基础之上继续预训练，两个实验的准确率是一样的，都很低。继续训练已经按照你说的加入了init_checkpoint参数，并且日志中也明确看到确实加载了，为什么结果会这么差？

#56 chenchengshuai closed 4 years ago
3
Pytorch版本的网盘失效了

#55 cqlyiyeshu opened 4 years ago
2
使用roberta-large去预测mask的位置，结果不可读

#54 yayaQAQ closed 4 years ago
1
关于预训练模型

#53 Foehnc closed 4 years ago
2
Roberta_l24_zh_base 和RoBERTa-zh-Large，有什么区别吗，还有有keras调用的样例吗，谢谢大佬了

#52 lx-rookie opened 4 years ago
3
RoBERTa-zh-Large百度网盘文件取消了

#51 jh-deng opened 4 years ago
6