Makes movie style screen grabs with all the noise and fuzz you'd expect of a movie clip.
Description
FAQ
Comments (9)
What's the suggested range of lora strength?
0.5 - 2.0
How did you train this? I am trying to train a cinematic light/color but always failed... Its works just with the base model. If i try to use to use on the distiled model, i have to push the values, but even doind that, it less weak effect!
I'm going to redo this one at some point. It was on ai toolkit
That's about all you need, and you also need ~50-200 steps per image, caption them with something like qwen3-vl and ask it to tell you exactly what is in the image, but to ignore the style
@anyMODE Thanks a lot i will try again and see if now works. Tried several times without any sucess
@anyMODE Its cooking and now its on 3.000 steps. Seems now its working. I dont get distortion on the distilled model and I will let cook until 6.000 steps. But Samples images are now starting to get distorted and weird... Is that normal? It happened to you too? Cause i will not use on base model, just distilled. And i am testing every time it saves (250 steps) seens on the right way!
@ronaldomirandah404 You can pretty much ignore the sample images. I don't bother letting it do them any more.
@anyMODE thanks! You were right again I will even cancel sample images. its getting there, thanks a lot! Sorry bother you again!
@ronaldomirandah404 No problem I like to help out
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.