LED icon indicating copy to clipboard operation
LED copied to clipboard

customized rawpy directory

Open wangchx67 opened this issue 1 year ago • 7 comments

Hi, thanks for your VERY impressive work! I have a problem when I installed customized rawpy of ELD. As said in ELD _" type pip install -e . in the customized rawpy directory", while I can not find the customized rawpy directory provided by ELD. Could you provide some suggestions?

wangchx67 avatar Jan 29 '24 14:01 wangchx67

If you installed the enviorment with our install.sh, our custimized rawpy was install in download/rawpy.

Or you mean the link to the customized rawpy: https://drive.google.com/drive/folders/1QoEhB1P-hNzAc4cRb7RdzyEKktexPVgy.

BTW, if our LED is helpful for your project, we welcome you to give us a star 🌟, which would be a great encouragement for our further updates!

Srameo avatar Jan 29 '24 15:01 Srameo

SURE, you deserve it! Thanks for your link, I have found it in file install.md.

wangchx67 avatar Jan 29 '24 15:01 wangchx67

Hi, I have an additional question about MIPI competition data. I find the checkpoint of NAFNet did not work for MIPI data as follows: image

But the Unet works: 1708441726953

Meanwhile, while finetuning NAFNet with the MIPI dataset and encountering a similar situation, do you have any suggestions regarding this?

wangchx67 avatar Feb 20 '24 15:02 wangchx67

Sorry that we are not allowed to discuss with the MIPI workshop 🥲

Srameo avatar Feb 20 '24 15:02 Srameo

😂😂Could you please clarify if different network architectures have a significant impact on performance for the SID or ELD datasets? I haven't come across corresponding tables.

wangchx67 avatar Feb 20 '24 17:02 wangchx67

Sorry for the late response, related performance on the SID dataset can be found in Tab. 8 in our updated paper.

Srameo avatar Mar 12 '24 06:03 Srameo

I also encountered the problem of fine-tuning on other datasets and found that the effects of NEFNet and Restormer are not as good as the effects of Unet. Could you please let me know if you have solved the problem? I would be grateful if you could reply to me.

xxxxyliu avatar Apr 21 '24 02:04 xxxxyliu