ZT

> ok > If there is a result, I will let you know as soon as possible > Your results look like they should be fine if you keep training....

Maybe train the U-Nets separately? Easier for debugging

>  hi > > > > 好的，如果有结果，我会尽快通知您，如果您继续训练，您的结果看起来应该没问题。尝试再训练 24 小时 您可以尝试将 lr 更改为 1e-5 > > > > > > 谢谢。如果你不介意，你能看看问题[#245吗](https://github.com/lucidrains/imagen-pytorch/issues/245)？也许设置有问题，使我的训练如此缓慢？ > > This is my image. It looks like...

@lucidrains Hello, I noticed that you have another similar repo https://github.com/lucidrains/video-diffusion-pytorch which supports co-training images and videos. However, it seems like this image-pytorch repo does not support this. I would...

@lucidrains what's the reason for disabling all the attention layers in the super-resolution U-Net? Does that help improve the image quality?

@lucidrains Thanks for the clear explanation :)

@HelloWorldLTY The > flat not found error is caused by applying numpy methods on csr_matrix. The following code should solve it: ``` if singlecell: # Convert csr_matrix to lil_matrix for...

Hi, I'm also experiencing this issue. I suspect it has something to do with RFD3 not interpreting the three-letter code correctly, hence building the wrong ligand geometry.