diadestiny

Results 6 issues of diadestiny

![image](https://github.com/youngyangyang04/Skiplist-CPP/assets/44188454/8f46cfdc-e136-4c9c-9702-508b8be35b55) 把TEST_COUNT改成30w, 压力测试会出现段错误Segmentation fault (core dumped),有朋友知道原因么

Thanks for your great work! The paper mentioned you adopted SytleGAN2 based on BasicSR(https://github.com/XPixelGroup/BasicSR) as pretrained model. However, It actually used the SytleGAN2 based on MMGeneration(mmedit/models/components/stylegan2/generator_discriminator.py) in your mmediting code....

Thank you very much for your work. When will you open source code?

if I hope to use fastsam in my code to process the batch tensor(shape: b, c, h, w), how to use it? If anyone can give the example code or...

您好,这是我用两张3090运行训练rain100H的validation的log,代码需要迭代70w次,我中途迭代到38w次,psnr是在25.5附近(暂未训练完),感觉难以复现到论文的31.65,想排查下原因(没有改动代码,不清楚log是否正常,感觉不太稳定),方便上传您训练的log么;另外,train.py代码中的dataset_ratio多卡训练参数默认是200,是否需要调整,会对结果产生很大的影响么。 ![UUZP`PQV_CTKAOJ(O3RC$WH](https://github.com/Algolzw/image-restoration-sde/assets/44188454/86d7c333-fe2f-42dc-bf02-55381c4eb9dc) 祝好,期待您的解答!

感谢您的工作! 有一个小问题,您对比的方法大部分是基于RESIDE数据集来训练的,直接用他们的pretrained model和您的model对你的话,是不是不算太严谨,就是是否考虑过用同样的gt(比如RESIDE)来基于您提到的合成pipeline进行在线合成训练 来进行对比。 还想请问下您公布的500张数据集如果换成其他真实干净图像,再基于您的数据合成pipeline进行训练,是否会对结果有影响?