It steers the model towards the style of gpt image 2. Improving details and lighting a lot, though the image might end up looking slightly sloppy. Play with lora strength to allow more style diversity :)This is trained without a reg dataset, so the trigger word is not needed, but helps ever so slightly.

STYLE > gpt-image-2

Training Settings: batch size 4, 1024px, ~114 images per epoch, optimizer is <a target="_blank" rel="ugc" href="https://github.com/Koratahiu/Advanced_Optimizers/tree/goemtric_wd">SinkSGD_adv</a>I did a few epochs at 1216px &amp; 1344px after the model stabilized to bring out the details.

gpt-image-2_anima-base1_v1.safetensors

Anima-base1 v1.0

Anima-pre3 v1.1

Anima-pre3 v1.0

creator: agile363274It steers the model towards the style of gpt image 2. Improving details and lighting a lot, though the image might end up looking slightly sloppy.This is trained without a reg dataset, so the trigger word is not needed, but helps ever so slightly.

Training Settings: batch size 4, 1024px, ~114 images per epoch, optimizer is SinkSGD_adv

I did a few epochs at 1216px & 1344px after the model stabilized to bring out the details.


Files with SHA256: 206a4d5ab78834de343f99523d34e18c4c0a2822621d053b29b9a2ef35ef6445