ShadowFormer

ShadowFormer copied to clipboard

RMSE or MAE

1

May I ask the author, why is the RMSE written in the evaluation index in the paper, but the MAE is calculated on the evaluation code?

Amy-214

Hi! Thanks for sharing the code. I'm interested in exploring it further and was wondering when the SRD pre-trained model weights, as mentioned in the original paper, will be available?...

irenesu2000

As the title. I didn't find the supplementary for the paper.

Pokerman8

ValueError: Since image dtype is floating point, you must specify the data_range parameter. Please read the documentation carefully (including the note). It is recommended that you always specify the data_range...

gejutaixiao22

Hi! Thanks for sharing the code. As described in the original paper, would like to know where I can access to the supplementary document. Thanks a lot.

FireWallDragonDarkFluid

SRD result is missing.

3

Hi. I'm doing an evaluation while looking at the paper, and I have a question. Among the result images for test data set, the image for SRD data set is...

DongHyun99

Hi! I'm wondered about your results. I have a question about processing big size images (2K, 4K). Are you going to create inference script which will process image by tiles...

heorhiikalaichev

ground truth image availability or workaround

1

Hi thank you for sharing such great work! I would like to fine tune the model on our data, but the ground truth i.e. shadow free images are not available...

pengzhao-life

I run into CUDA OOM errors for 1024x1024 even for a batch size of 1, any suggestions how we might be able to fit it to the memory?

TolgaAktas

Segmentation fault after 14 epochs training

1

When I try to use **custom dataset** for training, **Segmentation fault** occurs **after completing 14 epochs training**. How can I fix it? **datase**t: custom data (training set: 475 , validation...

hnrna