Now avaiable for 1.5 aswell ! Please check the metadatas images in the version section for prompts.
I'm happy to share with you my embedding carefully and progressively trained with a very high vibrant style and pop art influenced.
Most of thoses generations are 20-30 steps on Euler a, don't push too much the steps otherwise SD will try to put too many details if you want clean compositions.
Trained on 2.1-768 SHA [e1542d5a]
A clean 1980 scene of a classy man in suit medium shot and wearing sunglasses with a (city) in the background at sunset, soft lighting, nostalgic, art by rzviceb6
A clean aesthetic scene of a cute woman with beautiful eyes, medium shot with a (city) in the background at sunset,art by rzviceb10, soft lighting, cinematic lighting, nostalgic,
A digital illustration of a {computer | guitar | retro television... } standing on a table, art by rzviceb6
Embeddings steps comparison 6 is 6k 10, 10k
A clean scene of a modern woman, medium shot with a (city) in the background at sunset, soft lighting, nostalgic, art by rzviceb6
"A clean 1980 scene of a retro car parked in front of an hotel, nostalgic, art by <rzvicebx>"
if you want less influence of the embedding use the brackets [] [rzvicebx] to lower the weight and / or add the token to the end of your prompt
reminder of some of good negatives :
poorly Rendered face,poorly drawn face,poor facial details,poorly drawn hands,poorly rendered hands,low resolution,bad composition, mutated body parts,blurry image,disfigured,oversaturated,bad anatomy,deformed body features
I would love to see what you got with it !