Peter

> 感谢你的回复。这个项目中Bing相关的内容对我帮助很大。**这是我在settings.js** 中设置代理的地方。为了方便观察，我删除了注释的内容，在实际代码中保持不变。**我在use-client.js** 中将它设置在与 clientOptions 相同的位置。******** > > > ``` > > chatGptClient: { > > modelOptions: { > > // You can override the model name and any other...

I deleted the distributed code, and then it can run normally。

There seems to be a problem with distributed running code. When I entered the training command, the program did not respond, and I had to delete the distributed code

> Actually, I also met this problem when I was using another version pytorch. It seems to only work for "1.9.0+cu111". Okay, I'll try later

ground truth  Generate results  

> > 谢谢！但是从官方网站来看，他们似乎能够通过单个 img + PTI 获得相当不错的结果。 > > inversion_compressed.mp4 > > 您使用的输入图像有耳朵遮挡，视频中的输入图像包含更完整的信息。此 repo 无法生成被遮挡的区域。 > > 你可以看到我使用我的 repo 生成的结果：[NVlabs/eg3d#28 (comment)](https://github.com/NVlabs/eg3d/issues/28#issuecomment-1159512947)， 这里是输入重新对齐的图像和输入相机参数： [01457.zip](https://github.com/NVlabs/eg3d/files/8947809/01457.zip)  **Why does the regenerated image look...

> 我为生成猫图片训练了 20,000 步。猫图片的质量在前 15,000 步中越来越好。但很难说在那之后图像质量是否还在提高。顺便说一句，在我的案例中，在前 4,000 步之后，损失并没有显着减少。 > > 结果还可以，但并不令人印象深刻。[您可以在https://drive.google.com/drive/folders/1-obvOanCKA0lRGeS-SFK6eiZmAVvgqhn?usp=sharing](https://drive.google.com/drive/folders/1-obvOanCKA0lRGeS-SFK6eiZmAVvgqhn?usp=sharing)查看它们[](https://drive.google.com/drive/folders/1-obvOanCKA0lRGeS-SFK6eiZmAVvgqhn?usp=sharing) I saw your generated results and felt that the picture was not very clear. I am using bedroom, and...

i get the same problem,Can you explain the reason，thank you

> 正如我所说的查看 EMA，与“模型”参数相比，指数移动平均线的参数变化非常缓慢。修复：训练（远）更长的时间。或者从具有“模型”重量而不是 ema 的模型中采样 I have tried both model and EMA model, and the results of both are noise points，There is also an opt model. I don't know what...

> I have encountered the same problem. Have you solved it? Thank you. I use guide - ddfm to train, and within the same training time, the guide - ddfm...