Lin Qi

Hi there, thanks for making such an interesting work, after using the code `>python main.py --do_train --lm_type gpt2 --field_ce --flatten --data_type card --output_dir ./ --data_root ./data/card/ ` and I get...

请问如何实现chat的形式

3

您好，感谢分享，想问一下如何实现连续问答呢？ 是通过每次把上一次的输入和输出都加入到下一次的输入中么？ 谢谢

关于继续训练（continue finetuning）

12

你好，感谢分享。 是否可以理解为，在垂类应用上，构造好预料，替换sh中的data_path，然后运行sh文件就能在你们分享的模型权重下继续进行训练，以得到一个应用于上述垂类语料中的模型呢？ 谢谢

Which method can generate weighted edges?

4

Hey, just wanna know which method can generate weighted edges for different datasets. I know lingam can do this, but it assume the linear relation between columns, so is there...

[BUG] Recover to the latest checkpoint if NaNs encountered during training of GluonTS models

2

Currently, GluonTS models completely fail if NaNs are encountered at any point of the training process. A better alternative would be to recover to the latest non-NaN checkpoint of the...