✨ Save 50% on Annual Plans ✨ Claim 6 months FREE →
एक स्थान पर सभी सर्वोत्तम AI मॉडल्स
नैनो केला, जीपीटी इमेज, फ्लक्स, सीड्रीम, सीडांस और अधिक। इसे अभी मुफ़्त में आज़माएँ!
a year ago
Anagene Ayers - New Krea Lora Test
सृजित a year ago · 2 टिप्पणियाँ· 0 लाइक
Uploaded
Anagene Ayers tilted her head, mesmerized by the sphere’s gentle glow. It pulsed with an inner light, reflecting the colors of the setting sun. Was it merely a trick of the light, or something more? A sense of wonder filled her as she continued to watch, unsure of what secrets it held.
these 3 are the first tests of a completely redone dataset that went back to the original SDXL model i created them all with, and then did an in-line krea upscale to 1536x1536, which seemed to help make them more realistic looking and increase a bit of detail, then changed a few settings and went with a lower learning rate and trained ALL the available resolutions, and this is the 1000 step result, which only took a short while, even the low 200 step results were recreating her nicely.
2 टिप्पणियाँ
बातचीत में शामिल हों
Tigapup @Tigapup
This one isn't published? 😯
She's captured so well and so gorgeous in this, I would definitely publish it! 😍🩷✨️
2026-03-04T13:34:33.842ZReply
सृजन सेटिंग्स
यह सृजन NightCafe पर नहीं बनाया गया था।
- मॉडल
Flux Krea Nunchaku & WanVAE 2x Upscale
- प्रॉम्प्ट्स
- using a specialized prompt sent to Qwen 3-8B to take the theme and instructions, a large number of random locations, then create a prompt with the character's lora, and this is the result: a close-up portrait of AnageneAyers, her white-orange ombré hair cascading in soft waves against the golden-hour ...
- वर्कफ़्लो
- ComfyUI used to gen the source, then ComfyUI again to fix anything using Flux Krea Nunchaku for any inpaints, then SDXL w/ ControlNet direct 4k upscale(meaning no tiling), OR using the new Wan VAE 2x Upscale, and finally GIMP to do any pre or post-fixes, as well as final color correction and enhancing, and signaturing. and newly added is using Gemma-3-12B in LM Studio to generate short stories about the images to give a touch of curiosity to them.
