Spaces:
Runtime error
Runtime error
How can I upload multiple videos of the same type and train them to have a consistent style?
#57
by
cherff
- opened
As a filmmaker, I have a lot of driving videos from various cities. I want to feed them all to AI so that in the future, I can simply input the text "driving scene" and it will generate a video of driving on city roads without me having to collect the footage myself.
Yes, it would be great to have a GUI that can help me upload videos for training! May I ask how you trained your videos and how you labeled them? How can I operate it?
You might be interested by this:
https://github.com/ExponentialML/Text-To-Video-Finetuning
wow, that's cool, i will try it😁 And two more question:
- are there any requirements for the size of the video? I saw in your example that it was written as 1024*576.
- plz look at the images, The videos I often capture are in groups, taken from different angles of the same car. Often, five cameras are mounted on a single car, capturing videos from the front, back, left, right, and top. These videos are easy to play and capture on XR devices. My question is, how can I train this type of video captured in groups? Thanks♪(･ω･)ﾉ
This is a very specific case, you might ask @cerspense about how to fine tune train a set of videos :)
