Model trained on liminal space pictures I took after applying a simple VHS filter on it, using https://civarchive.com/models/13941/epinoiseoffset in combination with this model works well!
Description
FAQ
Comments (13)
Great model, however I think it would be more versatile if it lacked the VHS filter, having the end user create the effect (if needed).
Thank you! I agree but that's the whole reason I made the model hahaha, maybe the next one will be more versatile, but we'll see, also I just remembered the app I took the images for the training data had effects applied in the camera app itself, I got confused between this training data and another set I have
That loses the point of the whole thing wouldn't it? :P
@semi_neural you could create a VHS overlay LORA im pretty sure, this does seem neat.
i wonder about the possibility of using blender renders in the training data
@olivetty Not really, as I could dial the VHS look however I want, having so me direct experience with recreating the effect
good thing they deleted their account since they didn't get the point of AI models being able to do things without post editing lol.
He was correct, you only need the space, not the filter. Otherwise you'll have different tones and grain in each image and no consistency, so you can't make a comic, video or pack of images, as all of them look different.
Yes, a smaller LORA would be great !
this would be more suited as a lora if possible please
I will try and see if I can make one! I'm pretty new to training so I don't really know how to do that but I'll def try and figure it out! :)
@semi_neural There are ways to extract the lora from the original training. I personally prefer the checkpoints though.
SlientHill Ghost style! Loving it!
I really really wish this was safetensor.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.