app.py

import torch
import gradio as gr
from PIL import image
import spaces
from diffusers import StableCascadeDecoderPipeline, StableCascadePriorPipeline

device = "cuda"
num_images_per_prompt = 1

prior = StableCascadePriorPipeline.from_pretrained("stabilityai/stable-cascade-prior", torch_dtype=torch.bfloat16).to(device)
decoder = StableCascadeDecoderPipeline.from_pretrained("stabilityai/stable-cascade",  torch_dtype=torch.float16).to(device)

prompt = "Anthropomorphic cat dressed as a pilot"
negative_prompt = ""

@spaces.GPU
def gen(prompt, negative, width, height):
    prior_output = prior(
        prompt=prompt,
        height=height,
        width=width,
        negative_prompt=negative,
        guidance_scale=4.0,
        num_images_per_prompt=num_images_per_prompt,
        num_inference_steps=20
    )
    decoder_output = decoder(
        image_embeddings=prior_output.image_embeddings.half(),
        prompt=prompt,
        negative_prompt=negative,
        guidance_scale=0.0,
        output_type="pil",
        num_inference_steps=10
    ).images
    return decoder_output