MM-RealSR

MM-RealSR copied to clipboard

MMRRDBNet和MMRRDBNet_test

3

你好，看完了工程我有一个疑惑，在mmrealsr_arch.py中有两个模型MMRRDBNet和MMRRDBNet_test，训练阶段使用MMRRDBNet，测试阶段使用MMRRDBNet_test，两个的区别在计算degree的部分，但是我看不懂为什么训练阶段需要计算一个min/max？

ai1361720220000

x2 Weights missing

1

Great work with this SR, I'm researching SR solutions and doing a deep comparison between MM-RealSR and other SoTa options but can't do a real comparison without the x2 model,...

ofirkris

New Super-Resolution Benchmarks

1

Hello, [MSU Graphics & Media Lab Video Group](https://videoprocessing.ai/) has recently launched two new Super-Resolution Benchmarks. * [Video Upscalers Benchmark: Quality Enhancement](https://videoprocessing.ai/benchmarks/video-upscalers.html) determines the best upscaling methods for increasing video resolution...

EvgeneyBogatyrev

training iterations

3

Hi, @MC-E. I really thank for your awesome work. I have some questions about training MM-RealSR. 1. In your paper, MM-RealSR was trained during 1000k iterations for phase 1 and...

sujyQ

Hi, I want to finetune mm-realsr on my own dataset, could you please provide the net_d model? Thank you!

universea

device

1

Hello, the author. Thank you very much for your work. I want to ask you how long you have been training and what kind of GPU you use

yxzhe-star

About the default_init_weights

1

Thanks for your great work! Currently, I try to reproduce your work. However, I meet some problem when I couldn't find the function `default_init_weights` . I would appreciate it if...

csguoh

Thanks for your great work! Currently, I try to reproduce your work. However, I meet some problem when I couldn't find the function `default_init_weights` . I would appreciate it if...

csguoh

Thanks for your great work! Currently, I try to reproduce your work. However, I meet some problem when I couldn't find the function `default_init_weights` . I would appreciate it if...

csguoh

Loss

2

Great job, I'm interested in it. But when I trained with this code, I found that l_ranking_blur and l_ranking_noise kept at 5*10-4 without decreasing. What's the reason? Could you help...

Madeline-hyh