EasyEdit

EasyEdit copied to clipboard

Editing non-QA knowledge, such as DUNE dataset.

1

Hi, Thank you for your valuable open source contributions! I would like to inquire about how to utilize your toolkit for editing reasoning and non-QA knowledge. For instance, in the...

quqxui

made it grammatically correct

ancelia06

MultimodalEditor Question

1

Thank you for your outstanding work on multimodal knowledge editing. However, when I was trying to test your work, I found the following line of code in the **Edit** method...

wxz2002

I tried to reproduce the result of MEND in gpt-j-6B and Llama-2-7b, but the ngram-entropy of gpt-j-6B is far below Llama-2-7b(gpt-j-6B around 350 vs Llama-2-7b around 550). Do you have...

jiqimaoke

关于wise.py

3

非常感谢工作的分享，很有启发！有两个问题想请教： 1. 观察到论文和代码中只复制了一个层的weight，请问有考虑过优化两个层的weight吗？比如` transformer.h[7].mlp.c_fc.weight`和`transformer.h[8].mlp.c_fc.weight`。如果想修改的话，主要在wise.py中修改吗？ 2. 请问有考虑过引入bias吗？联合更新weight和bias

zhengguanyu

jiangziyan-693

GRACE sequential edit result

13

Hello, I am now playing with GRACE on llama2 recently but I've noticed a significant difference in the results between the default settings and the WISE paper's result. Could you...

SXxinxiaosong

Editing Llama-7b with MEMIT on multiple GPUs: Expected all tensors to be on the same device

2

Thanks for this great repo. I'm trying to edit Llama-7b with the MEMIT algorithm. With two GPUs, I get the following error: ``` File "easyeditor/editors/editor.py", line 182, in edit return...

fateme-hshm96

Question about Evaluation implement details

1

Hi, Thanks a lot for your great work! Recently I encounter the same question as #114 , but I still don't understand the logic behind your reply. I quote: >...

asdfo123

is cuda inevitable?[MacOS]

2

York-lap