This is a simple concept LoRA created for the Minimax H3 contest. This LoRA was trained on a couple of concepts:
Moving through wilderness foliage
Moving through dead trees, wooden sticks.
Moving through a fast moving road.
Filling a cup to the rim. Didn't work as intended, I didn't expect for the training step to trim all videos to 5 seconds. Unfortunately, I don't have enough Buzz to attempt to train a new version. Also, generating the videos for the showcase post have drained my Buzz even more.
Description
First version with 4 simple concepts.
Comments (2)
I think I understand what this model does, it's like a more photorealistic, sort of natural/cinematic angled camera shot of people moving through specific places?
I wouldn't call it cinematic, more like amateur shot. To be honest, this was my first H3 LoRA and I had no idea if it was going to work at all. Thanks for trying it.