Yuxuan Zhang

没有在模型实现的时候做这个工作，应该会报不支持吧，模型实现代码没写

now it is support diffuser framework using 4090

Of course, it is supported, but this requires you to implement it yourself. You can use the doubled channels to store the image lantent as the condition for training.

My understanding is that you want to make a transition between two videos, with the input and output corresponding to the last frame of the first video and the first...

Given this, a good suggestion would be to fine-tune the model for downstream tasks. This would require some changes to the model structure, but not much, as it's for a...

Yes, the I2V model is trained by doubling the channels of the T2V extension.

Yes, then expanded to 16 channels.

nop, can you share the log?

不正常，这个scale也不正常，你的数据集体量是 这个看上去是没有任何报错都是跳过了？

Did everyone skip all the steps? @kyrie111 @TianxingWu, because skipping the first few steps and then continuing with normal training, the loss reduction is a normal phenomenon. The first few...