enable_lora

[ErfeiCui]

I saw in the quick start section that the sample code includes enable_lora=True, but I noticed that Editscore's initialization function doesn't have this parameter, which causes an error during execution.

[Luciennnnnnn]

Hi, thank you for reporting this issue. We’ve cleaned up the code, so enable_lora is no longer required — you can simply omit it. The documentation has been updated accordingly.

[ErfeiCui]

Thank you for your reply! However, when I used the sample code, I encountered the issue shown in the image. What could be the reason for this?
Additionally, the result is as follows: result: {'prompt_following': 7.5, 'consistency': 5.0, 'perceptual_quality': 8.5, 'overall': 6.519202405202649}, and no 'final_score' was found.

[Luciennnnnnn]

Apologies for the confusion. There is no final_score key; the correct key is overall. This was a typo caused by LLM rewriting.

As for the Lora-related warning, it's unusual. I’ll investigate it soon. Could you please share your environment details?

[ErfeiCui]

Thanks, my main environment is:
torch 2.5.1+cu121
torchaudio 2.5.1+cu121
torchvision 0.20.1+cu121
transformers 4.49.0
editscore 0.2
vllm 0.7.2
Python 3.11.12

[Luciennnnnnn]

Did you just execute the example code from the README? If not, could you also share the code you ran?

[Luciennnnnnn]

I confirmed that I can successfully run example code.

[ErfeiCui]

I just execute the example code:
'https://github.com/VectorSpaceLab/EditScore?tab=readme-ov-file#-usage-example'
I will create the new environment, Could you share your environment details? Thanks

[Luciennnnnnn]

My environment is created exclusively following installation guidance, here is resulting environment:

torch 2.7.1+cu126
torchaudio 2.7.1
torchvision 0.22.1+cu126

transformers 4.57.1
editscore 0.2
Python 3.12.0

Did you test with vllm version?