Textual inversion embedding for SD2.0 and 768-v-ema.ckpt
Converts even simple promts into landscape paintings
it was trained with no people in the dataset, but it can also work for them
all examples are made with 20 steps, DPM++ 2M Karras and CFG 7 and 768x768 resolution, no other promts, no negatives
moon station, white block buildings, foreign planet, painted_landscape:
asteroid impact, huge sun, rocky landscape, painted_landscape:
foreign science fiction planet, weird tangled trees, purple plants:
black swamp, witch hut, mud water, painted_landscape:
small town marketplace, store and shop, medieval buildings, painted_landscape:
demon devil black stone tower, hellfire landscape, bone trees, painted_landscape:
vulcano, eruption, lava, painted_landscape:
mega cyberpunk city, neon lights, skyscraper, painted_landscape:
castle, high mountains, elvish architecture, painted_landscape:
portrait of a woman, red long hair, long black coat, painted_landscape:
portrait of a woman, top bun hair, frilly blue dress, painted_landscape:
portrait of a old man, grey bushy beard, catching a fish, painted_landscape:
