Morizeyao/GPT2-Chinese issues and pull requests

#43 - RuntimeError: Creating MTGP constants failed.

Issue - State: open - Opened by leemengtw about 5 years ago - 14 comments

#43 - RuntimeError: Creating MTGP constants failed.

Issue - State: open - Opened by leemengtw about 5 years ago - 14 comments

#42 - 句子过长导致的索引错误（Too long a sentence leads to an index error）

Issue - State: closed - Opened by lioyou about 5 years ago - 3 comments

#42 - 句子过长导致的索引错误（Too long a sentence leads to an index error）

Issue - State: closed - Opened by lioyou about 5 years ago - 3 comments

#41 - 增加sentencepiece支持

Pull Request - State: closed - Opened by kangzhonghua about 5 years ago - 1 comment

#40 - 添加自定义文件类型及数据源支持

Pull Request - State: closed - Opened by lioyou about 5 years ago - 1 comment

#40 - 添加自定义文件类型及数据源支持

Pull Request - State: closed - Opened by lioyou about 5 years ago - 1 comment

#39 - 用CPU执行train_single训练，报这个错误RuntimeError: index out of range

Issue - State: closed - Opened by leedaga about 5 years ago - 2 comments

#39 - 用CPU执行train_single训练，报这个错误RuntimeError: index out of range

Issue - State: closed - Opened by leedaga about 5 years ago - 2 comments

#38 - argparse中参数gradient_accumulation类型错误

Issue - State: closed - Opened by xinfeng1i about 5 years ago - 3 comments

#37 - where can download the trained model

Issue - State: closed - Opened by molsheim about 5 years ago - 1 comment

#37 - where can download the trained model

Issue - State: closed - Opened by molsheim about 5 years ago - 1 comment

#36 - 生成的古诗中包含大量UNK

Issue - State: closed - Opened by lijun20 about 5 years ago - 2 comments

#35 - Fail to run train_single

Issue - State: closed - Opened by diansheng about 5 years ago - 1 comment

#34 - Undefined name 'running_loss' in ./train_single.py

Issue - State: closed - Opened by cclauss about 5 years ago - 1 comment

#34 - Undefined name 'running_loss' in ./train_single.py

Issue - State: closed - Opened by cclauss about 5 years ago - 1 comment

#33 - Word branch

Pull Request - State: closed - Opened by fengzuo97 about 5 years ago - 4 comments

#32 - 多gpu混合精度训练的问题

Issue - State: closed - Opened by fengzuo97 about 5 years ago - 5 comments

#32 - 多gpu混合精度训练的问题

Issue - State: closed - Opened by fengzuo97 about 5 years ago - 5 comments

#30 - 较大规模训练后自由生成的文本。模型参数约80M。机器为四个2080Ti，训练步数140万步，语料3.4G，Batch Size 8。

Issue - State: closed - Opened by willshion about 5 years ago

#29 - 请问能否支持非续写模式的generate?

Issue - State: closed - Opened by lexmen318 about 5 years ago - 6 comments

#29 - 请问能否支持非续写模式的generate?

Issue - State: closed - Opened by lexmen318 about 5 years ago - 6 comments

#28 - nothing

Issue - State: closed - Opened by ashora about 5 years ago - 1 comment

#28 - nothing

Issue - State: closed - Opened by ashora about 5 years ago - 1 comment

#27 - 可否有监督

Issue - State: closed - Opened by WeiliangGuo about 5 years ago - 1 comment

#27 - 可否有监督

Issue - State: closed - Opened by WeiliangGuo about 5 years ago - 1 comment

#26 - 是否保存optimizer的参数？

Issue - State: closed - Opened by fengzuo97 about 5 years ago - 1 comment

#25 - cannot import name 'clean_up_tokenization'

Issue - State: closed - Opened by leemengtw about 5 years ago - 6 comments

#24 - 训练成果分享与一点提问

Issue - State: closed - Opened by chiangandy about 5 years ago - 17 comments

#23 - #分词训练格式

Issue - State: closed - Opened by kangkang61 about 5 years ago - 4 comments

#22 - 训练数据格式问题

Issue - State: closed - Opened by huaxiaohua about 5 years ago - 3 comments

#21 - 环境使用内存的问题

Issue - State: closed - Opened by chiangandy about 5 years ago - 4 comments

#20 - 软件包安装问题

Issue - State: closed - Opened by chiangandy about 5 years ago - 2 comments

#19 - #loss每个num_piece会先增加再减

Issue - State: closed - Opened by cyl250 about 5 years ago - 1 comment

#18 - 请问词表有预训练权重么

Issue - State: closed - Opened by ZhaoyueSun about 5 years ago - 2 comments

#17 - #局部收敛问题

Issue - State: closed - Opened by cyl250 about 5 years ago - 6 comments

#16 - 求教一般训练几个epoch？

Issue - State: closed - Opened by ZhaoyueSun about 5 years ago - 3 comments

#15 - 训练的loss或者ppl？使用分字和分词的效果对比？

Issue - State: closed - Opened by fengzuo97 about 5 years ago - 8 comments

#14 - 硬件

Issue - State: closed - Opened by nlpofwhat over 5 years ago - 2 comments

#13 - add some samples with fixed genre

Pull Request - State: closed - Opened by JamesHujy over 5 years ago

#12 - 请教只训练语言模型

Issue - State: closed - Opened by kangkang61 over 5 years ago - 1 comment

#11 - 请问是从预训练模型开始训练的吗？

Issue - State: closed - Opened by Dongfeng-He over 5 years ago - 2 comments

#10 - poem generation samples

Pull Request - State: closed - Opened by JamesHujy over 5 years ago

#9 - 请教体育新闻的语料数量

Issue - State: closed - Opened by chiangandy over 5 years ago - 1 comment

#8 - 正体中文支援的问题

Issue - State: closed - Opened by chiangandy over 5 years ago - 1 comment

#7 - 训练好的模型移至CPU上执行

Issue - State: closed - Opened by chiangandy over 5 years ago - 1 comment

#6 - 文本学习效果请教

Issue - State: closed - Opened by chiangandy over 5 years ago - 2 comments

#5 - 语料处理

Issue - State: closed - Opened by goingcoder over 5 years ago - 1 comment

#4 - Need help with gradient accumulation implementation.

Issue - State: closed - Opened by Morizeyao over 5 years ago - 1 comment
Labels: enhancement, help wanted, good first issue

GitHub / Morizeyao/GPT2-Chinese issues and pull requests