Please ❤️ this model and leave a review if you like it, it makes a massive difference!
Training this model took countless days and hundreds of dollars; your support is more than appreciated! Special thanks to @HailoKnight , @pixelworld_ai and @Temporarium for their massive support.
BriXL is a fine-tuned model with a special focus on science fiction and cinematic vibes. Its imperfection is what makes it shine! It has been trained on a lot of my original content as well as cinematic materials.
It is extremely good at making mechanical objects and giving a cinematic lighting without too much contrast or abusive color vibrance while giving you style flexibility from cartoon-ish to full photorealistic scenes.
While this is a first release, BriXL is a constant work in progress for as long as I got the funds to keep the training running. If you enjoy the model, please leave a review, if you don't like it, please tell me why!
I am fine with merges as long as you credit me
Description
FAQ
Comments (3)
Sadly I have to wait a bit before I can use this model as it seems to eat more vram than my old 2070 possesses. Here are the prompts I was using:
--P:Cinematic film still a concert stage with a band performing on it. focus on the singer, The singer is in forefront, a foot resting on a low horizontal speaker turned toward the stage, wearing cammo fatigues, combat shoes, a sleeveless tactical vest, aviator glasses, his hair cut in a brown mohawk, facial hair, tatoos on his arms
top quality, highres, highly detailed, photorealist, volumetric clouds, volumetric lighting, caustics, bokeh . shallow depth of field, vignette, highly detailed, high budget, bokeh, cinemascope, moody, epic, gorgeous, film grain, grainy
--N:anime, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, bad-hands-5, easynegative, By bad artist -neg, Unspeakable-Horrors-Composition-4v, (inaccurate limb:1.2), bad composition, inaccurate eyes, extra digit, fewer digits, (extra arms:1.2), aircraft, spacecraft
20 steps , with recommanded VAE, latent upscaler at 1
Nice.
Fyi, v4 doesn't seem to bode well with Add Details XL (https://civitai.com/api/download/models/135867), all images with it enabled comes out either blank or with abstract pixel artifacts.
Used weight at 0.75
Thanks Ze! Will keep an eye out for that! Ultimately I'd rather have a decent amount of details from the checkpoint itself without having to resort to LoRAs but I'll give it a test on v5!
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.