I don't understand how to use it

[schaoyang340]

This project is great and I used it quickly, but I see that the output is logging model parameters. I think I may have understood something wrong. I put a blurred picture in the test data. I thought the output would be a denoised picture. So, to achieve my expectation, should I use the DDNM mentioned in the previous question?

[yuanzhi-zhu]

Sorry I don't really understand your question. What task are you doing?

[schaoyang340]

Sorry I don't really understand your question. What task are you doing?

My task is to add noise to an image and then remove it. I hope to use this project to complete the task of denoising. But I didn't understand how to use this project to achieve this task. I put a noise-enhanced picture under testsets. I thought a denoised picture would be output, but I only saw the log record. What should I do to make it output denoised images?

[yuanzhi-zhu]

Sorry for that I did not implement the code for denoising. As you've noticed from the previous issue, you can refer to DDNM for denoising task. Best,

[schaoyang340]

Sorry for that I did not implement the code for denoising. As you've noticed from the previous issue, you can refer to DDNM for denoising task. Best,

I wish to use DnCNN based denoising method. I saw image denoising based on DnCNN in the paper, so I found this project of your laboratory: https://github.com/cszn/DnCNN. But this project lacked model files, then I found another project, DPIR: https://github.com/cszn/DPIR. Unfortunately, it's also missing model files. In the Google download link given under the model_zoo folder, the files drunet_gray.pth and drunet_color.pth are missing. The Tencent Weiyun link failed before my eyes. I'm not sure if there are two missing files in this link. The DPIR project appears to be able to output denoised images. I hope my understanding is correct. If the DPIR project can indeed output denoised images, can you contact the author? Please ask this author to add these two model files as he doesn't seem to be logged in frequently enough to notice this issue.

[schaoyang340]

The translation may be a bit strange, please forgive me, I am a Chinese college student and was assigned a task that I don't quite understand.ヾ(•ω•`)o

[yuanzhi-zhu]

You can take a look at https://github.com/cszn/KAIR, where you should find most model checkpoints available

[schaoyang340]

Perfect. I used dncnn in KAIR. It meets my requirements and outputs denoised pictures. Thank you very much.

[ruiyiiii]

Perfect. I used dncnn in KAIR. It meets my requirements and outputs denoised pictures. Thank you very much.

Can you help me? I'm having trouble running the code。

[schaoyang340]

完美的。我在KAIR中使用了dncnn。它满足我的要求并输出去噪的图片。非常感谢。

你能帮助我吗？我在运行代码时遇到问题。

发生了什么问题，让我试试

[schaoyang340]

使用说明.pdf Instructions for use.pdf

If you want to quickly use the denoising method in the KAIR project, I have attached an instruction manual compiled for my classmates who use computers less. My classmate needed this noise removal task, but he didn't know much about its use, so I made an explanation

[ruiyiiii]

完美的。我在KAIR中使用了dncnn。它满足我的要求并输出去噪的图片。非常感谢。

你能帮助我吗？我在运行代码时遇到问题。

发生了什么问题，让我试试

请问ffhq_10m数据怎么下载呢？文档里的链接我打开不了，用bash download.sh只下载下来了256x256_diffusion_uncond.pt(ILSVRC 2012 subset of ImageNet)这个

[schaoyang340]

完美的。我在KAIR中使用了dncnn。它满足我的要求并输出去噪的图片。非常感谢。

你能帮助我吗？我在运行代码时遇到问题。

发生了什么问题，让我试试

请问ffhq_10m数据怎么下载呢？文档里的链接我打开不了，用bash download.sh只下载下来了256x256_diffusion_uncond.pt(ILSVRC 2012 subset of ImageNet)这个

啊？打不开链接，https://drive.google.com/drive/folders/1jElnRoFv7b31fG0v6pTSQkelbSX3xGZh，这个链接打不开是吗，这种情况是网络问题吧，你需要fanqiang，你既然可以浏览GitHub的话，在这上面自行了解怎么操作就好

[ruiyiiii]

对不起，我没有实现去噪代码。正如您在上一期中注意到的，您可以参考 DDNM 进行降噪任务。最好

我希望使用基于 DnCNN 的去噪方法。我在论文中看到了基于DnCNN的图像去噪，所以我找到了你们实验室的这个项目：https://github.com/cszn/DnCNN。但是这个项目缺少模型文件，然后我找到了另一个项目，DPIR：https://github.com/cszn/DPIR。不幸的是，它也缺少模型文件。在 model_zoo 文件夹下给出的 Google 下载链接中，缺少文件 drunet_gray.pth 和 drunet_color.pth。腾讯微云链接在我眼前失败了。我不确定此链接中是否缺少两个文件。 DPIR项目似乎能够输出去噪图像。我希望我的理解是正确的。如果DPIR项目确实可以输出去噪图像，您能联系作者吗？请要求该作者添加这两个模型文件，因为他似乎登录的频率不够高，无法注意到此问题。

你好,我的需求是用一种基于训练好的去噪扩散模型求解图像反问题(比如图像恢复)，这篇文章是做的这个吗，我有一点懵

[schaoyang340]

对不起，我没有实现去噪代码。正如您在上一期中注意到的，您可以参考 DDNM 进行降噪任务。最好

我希望使用基于 DnCNN 的去噪方法。我在论文中看到了基于DnCNN的图像去噪，所以我找到了你们实验室的这个项目：https://github.com/cszn/DnCNN。但是这个项目缺少模型文件，然后我找到了另一个项目，DPIR：https://github.com/cszn/DPIR。不幸的是，它也缺少模型文件。在 model_zoo 文件夹下给出的 Google 下载链接中，缺少文件 drunet_gray.pth 和 drunet_color.pth。腾讯微云链接在我眼前失败了。我不确定此链接中是否缺少两个文件。 DPIR项目似乎能够输出去噪图像。我希望我的理解是正确的。如果DPIR项目确实可以输出去噪图像，您能联系作者吗？请要求该作者添加这两个模型文件，因为他似乎登录的频率不够高，无法注意到此问题。

你好,我的需求是用一种基于训练好的去噪扩散模型求解图像反问题(比如图像恢复)，这篇文章是做的这个吗，我有一点懵

DiffPIR项目是基于去噪扩散模型的图像恢复任务，应该是符合你的要求