skykiseki

why is python3.6 not longer supported?

1

3.6 is a stable version of python, which remains a lot users. It's a little bit sad that it doesn't support 3.6 from v0.0.65

能否添加更多的模型使用手册，让新用户更友好呢？

2

只是提个建议，从苏神那边过来知道的，觉得openbuddy的潜力很大，但是文档之类的实在是不够丰富，新用户进来时候的体验是很不友好的，希望能优化。 补充一句：大佬们牛逼

accelerate 不支持对抗训练

1

提问时请尽可能提供如下信息： 基本信息 你使用的操作系统: ubuntu20.04 你使用的Python版本: 3.8.10 你使用的Pytorch版本: 1.10.2 你使用的bert4torch版本: 0.2.7.post2 你加载的预训练模型: bert-base ### 核心代码 ```python adversarial_train = AdversarialTraining('fgm') model, optimizer, train_dataloader, eval_dataloader, lr_scheduler, adversarial_train = accelerator.prepare( model, optimizer, train_dataloader, eval_dataloader,...

Any experiments in the future?

Seems like it doesn't work on my jupyter

1

Have restarted my jupyter server for times, but it still doesn't work on my jupyter The inspection is as follow, seems like some network problems happen, why? 

keras信徒表示在这里学（chao）到了很多，来这里跟大佬发个感谢，好像代码风格看到了苏老师（雾）

pandas低版本无法加载情感词典

1

python 3.7 pandas 1.1.5 import cntext as ct concreteness_df = ct.load_pkl_dict('concreteness.pkl') 按readme运行后 Can't get attribute 'new_block' on 升级pandas至最新版本1.3.5后解决，是不是安装包应该约束一下pandas版本？

It also combine qlora of longlora, what's the difference between longQLora and longlora+qlora? Model url as follow: https://huggingface.co/Yukang/LongAlpaca-7B-qlora-weights

Is there any comparison of the effects related to token attention? For example, compare with page attention

2

If there's a paper or other proof, that would be even better.