Zekun Xi comments

Results 10 comments of


                                            Zekun Xi

PMET Llama No Difference In Output

Thank you for your attention. We are currently busy with the NeurIPS submission, we will address this matter by the end of the month.

Suggestions for PMET Method on Llama

Thank you very much for your interest in EasyEdit. We apologize for our limited availability as we are currently busy with the nips submission deadline. We will focus on optimization...

line 118 in ft_main.py

Thank you very much for your advice; we have updated the code. Wishing you a pleasant life!

Editing non-QA knowledge, such as DUNE dataset.

Thank you for your attention. We plan to support this method in February 2024.

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

Thank you very much for your interest in EasyEdit. We will debug and address this issue soon. If you are in a hurry, you can manually set the tensor to...

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

Dear [fateme-hshm96](https://github.com/fateme-hshm96): Hello, you can try our latest code. I tested it locally with two A800 GPUs, and it runs smoothly with `model_parallel: true` enabled. EasyEdit Team

GRACE's performance collapes on LLAMA-3-8B

This situation seems quite strange. You can try using LLaMA2 to see if the same issue occurs. If it doesn’t, it might indicate that there are some compatibility issues between...

GRACE sequential edit result

Thank you for your interest in EasyEdit. We will reproduce this result and resolve the issue in the near future.

GRACE sequential edit result

vanilla_generation设置为Ture是因为GRACE使用了一个adapter，调用model.generate()会将key_id 设置为-1 ，以确保token_to_edit 能被正确设置。

GRACE sequential edit result

嗨👋，如果您还有进一步的问题，能否加一下微信，进行更高效率的沟通 xi1786594371