ShadowDiffusion

ShadowDiffusion copied to clipboard

Training iteration, epochs and batch_size for reproduce

1

I really interested in your work, and tried to reproduce your score. But, in paper, you illustrate training epochs is 1000 without batch size denoted, but i cant reproduce it....

cloor

training time issue

1

I appreciate your research efforts. I have a keen interest in your work. I would like to retrain the model following your paper, but training for 1000 epochs on the...

cloor

https://github.com/GuoLanqing/ShadowDiffusion/blob/a7e7aff3028e7d7e89be0f3d3d46479dbcea00dd/README.md?plain=1#L33C96-L33C105 Hi, Lanqing @GuoLanqing, thanks for your great work! I found an empty link when trying to download the 'ISTD'. Was it a typo?

MqLeet

About Predict

1

If now I only have a shadow image, how can the model predict and generate the shadow-free image ? Does the model include a mask detector?

finnn00

FireWallDragonDarkFluid

Good day, I have looked at the supplementary materials and I see your low light enhancement results. How can I use the pretrained model for low light enhancement? Thank you

savvaki

23-12-11 08:20:46.337 - INFO: Model [DDPM] is created. 23-12-11 08:20:46.337 - INFO: Initial Model Finished 23-12-11 08:21:49.840 - INFO: l_pix: 3.2691e-01 23-12-11 08:22:51.153 - INFO: l_pix: 2.1998e-01 23-12-11 08:23:46.322 -...

gejutaixiao22

你好，论文中说该方法可以应用在低光照增强任务中，低光照增强应该无需掩码，本文方法应该如何应用。

MrWan001

Thank you for your contribution in this field! Would you mind sharing the model pre-trained using Uformer? Looking forward to your reply！

hosophie

Hello, your work is excellent: Test question: I don't quite understand the process after the training ends. After my training, I generated many. pth files. Can these pth files be...

122443