SCTNet
SCTNet copied to clipboard
[ICCV 23] Official implementation of Alignment-free HDR Deghosting with Sematnics Consistent Transformer (Code & Dataset)
你好, 首先,感谢您精彩的工作 我在尝试复现模型的时候,出现了显存报错。debug时发现在G-SAB模块中,tensor的尺寸过大(150万),导致显存溢出。但是我并未修改您的模型。请问是什么问题呢? 期待您的回复!
Thanks for your excellent work, can you offer the code used to get HDR ground truth from LDR inputs?
I'm trying to work with the repository from Nonu116 but even testing with the pretrained model gives me bad results. sorry for the off topic issue
I use 2 NVIDIA A100 80GB PCIe, 160GB GPU memory total, but I still got CUDA out of memory error. I didnot change ur code, just git clone, pip install...
How ur dataset has test images of 1496 x 1000 size, but your results benchmark has 1500 x 1000 of predicted image ?  
Thank you for your excellent work. I'd like to ask how many GPUs you used for training the model and how long the training process took?