This model is pretrained on a collection of ZX-Spectrum images from zxart.ee

It generates images like this:

Usage example:

from diffusers import AsymmetricAutoencoderKL, DiffusionPipeline
import torch

PROMPT = "Blue racing car"
SEED = 123
STEPS = 20

pipe = DiffusionPipeline.from_pretrained("shadowlamer/sd-zxspectrum-model-256",
                                         safety_checker=None, requires_safety_checker=False)
pipe.vae = AsymmetricAutoencoderKL.from_pretrained("cross-attention/asymmetric-autoencoder-kl-x-2")
generator = torch.Generator("cpu").manual_seed(SEED)
image = pipe(PROMPT,
             height=192,
             width=256,
             num_inference_steps=STEPS,
             generator=generator,
             ).images[0]
image.save("result.png")