video prediction

[XXinLiu]

Hi, I would like to ask you two questions about the code you have written:
1: What is the structure of the text-video dataset
2: How to embed the first frame of the video during the video prediction task。
Looking forward to hearing from you!

[LiQiiiii]

Hi , have you implemented the code related to the video prediction setting? I'm also confused about it.

[XXinLiu]

Hi, it's an honor to receive your reply, my name is Liu Xin, a doctoral student from China. I tried changing the text-guided video generation in your code to text+image video prediction by encoding the images. I hope my explorations can be helpful and look forward to hearing from you. 在 2023-08-04 14:54:15，"Ritchie" ***@***.***> 写道： Hi , have you implemented the code related to the video prediction setting? I'm also confused about it. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>

[chpk]

Hi,

I have an issue similar to yours's. I have trained the model on a custom data, the milestones/checkpoints are saved in "./results" folder. The training code was straight forward, but during inference, i am not able to figure out how to infer on a custom input .gif using the saved checkpoint/milestones present in the "./results" folder. Any lead in this regard?

Thanks in advance,