Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / yangjianxin1/cpm issues and pull requests
#27 - 有关问答对话
Issue -
State: open - Opened by leizhu1989 almost 2 years ago
#26 - Crash on "poch_mean_loss = total_loss / len(train_dataloader)"
Issue -
State: open - Opened by cxsjabc almost 2 years ago
#25 - 出文比较慢,有解法没?
Issue -
State: open - Opened by wysstartgo almost 2 years ago
#24 - sentencepiece 模型
Issue -
State: open - Opened by TheHonestBob over 2 years ago
#23 - 关于生成时的加速
Issue -
State: open - Opened by syspider over 2 years ago
#22 - 微调OOM问题
Issue -
State: closed - Opened by TheHonestBob over 2 years ago
#21 - 关于模型输入label的问题
Issue -
State: open - Opened by shawroad over 2 years ago
- 2 comments
#20 - 如何提高解码速度?
Issue -
State: open - Opened by outstandingJie over 2 years ago
#19 - 如何提高生成的句子是否相关性?
Issue -
State: open - Opened by outstandingJie over 2 years ago
#18 - 大神,训练和生成都报错,请问这是什么问题呢
Issue -
State: open - Opened by guangh01 over 2 years ago
- 3 comments
#17 - 生成文本太慢
Issue -
State: open - Opened by Biaocsu almost 3 years ago
- 1 comment
#16 - 请问这个预训练模型可以有办法转化为GPT2直接使用的模型么?
Issue -
State: open - Opened by liruixue about 3 years ago
#15 - 用预加载模型,生成文本的时间比较长,有什么优化之道呢?
Issue -
State: open - Opened by liruixue about 3 years ago
- 1 comment
#14 - 训练数据
Issue -
State: open - Opened by shenshengke701 over 3 years ago
- 3 comments
#13 - train data
Issue -
State: closed - Opened by shenshengke701 over 3 years ago
#12 - 训练完后,不会生成一个总的模型文件,而是每轮训练完成后自动生成一个模型文件,这个最后在生成文章的时候,该选择哪一轮的模型作为生成模型
Issue -
State: open - Opened by guangh01 over 3 years ago
- 2 comments
#11 - 您好,我从CPM官网下载了pt格式的模型文件,您能分享下转换为bin格式的方法吗?感谢!
Issue -
State: open - Opened by kavinwow100 over 3 years ago
- 4 comments
#10 - 可以分享一下Finetune的过程吗?
Issue -
State: closed - Opened by xiewenqian over 3 years ago
- 1 comment
#9 - Preprocess.py 的 53-63行感觉有点问题
Issue -
State: open - Opened by MaNing1924382115 over 3 years ago
#8 - 如果重新训练模型,输入的tokens的向量表示的随意初始化的还是。。。?
Issue -
State: open - Opened by MaNing1924382115 over 3 years ago
#7 - 我想问一下,cmp109M那个模型在哪下载呢,去源码哪里没找到
Issue -
State: closed - Opened by guangh01 over 3 years ago
#6 - CMP源码申请下载为什么只有large参数模型,为什么没有CPM-small&CPM-medium,你用的small参数模型在哪下载的
Issue -
State: open - Opened by MaNing1924382115 over 3 years ago
- 4 comments
#6 - CMP源码申请下载为什么只有large参数模型,为什么没有CPM-small&CPM-medium,你用的small参数模型在哪下载的
Issue -
State: open - Opened by MaNing1924382115 over 3 years ago
- 4 comments
#5 - 申请下载的只有large,你用的small在哪下载的
Issue -
State: closed - Opened by MaNing1924382115 over 3 years ago
- 1 comment
#4 - 在使用train.py训练模型的时候,读取预训练模型,然后进行初始化,这个好像初始化不了
Issue -
State: open - Opened by guangh01 over 3 years ago
- 6 comments
#3 - 请问pretrained_model怎么加载原版CPM的参数呢?
Issue -
State: open - Opened by JiamingUWU over 3 years ago
- 1 comment
#2 - 是否使用CPM预训练模型
Issue -
State: closed - Opened by bupticybee over 3 years ago
- 3 comments
#1 - 现在不用tensorboard观察数据啦?
Issue -
State: open - Opened by kavinwow100 over 3 years ago