app.py

from diffusers import AutoPipelineForText2Image
import torch
import random
import os
import gradio as gr

hf_token = os.getenv("HF_TOKEN")
model_id = int(os.getenv("Model"))
nsfw_filter = int(os.getenv("Safe"))

#stable-diffusion-xl-base-1.0  0 - base model
#Colossus_Project_XL           1 - better people
#AlbedoBaseXL_v11              2 - realistic
#JuggernautXL_v7               3 - better faces
#RealVisXL_V2.0                4 - better photorealism

model_url_list = ["stabilityai/stable-diffusion-xl-base-1.0/blob/main/sd_xl_base_1.0.safetensors",
                 "Krebzonide/Colossus_Project_XL/blob/main/colossusProjectXLSFW_v202BakedVAE.safetensors",
                 "Krebzonide/AlbedoBaseXL_v11/blob/main/albedobaseXL_v11.safetensors",
                 "Krebzonide/JuggernautXL_version5/blob/main/juggernautXL_v7Rundiffusion.safetensors",
                 "SG161222/RealVisXL_V2.0/blob/main/RealVisXL_V2.0.safetensors",
                 "Krebzonide/AcornIsSpinning_acornXLV1/blob/main/acornIsSpinning_acornxlV1.safetensors"]

naughtyWords = ["nude", "nsfw", "naked", "porn", "boob", "tit", "nipple", "vagina", "pussy", "panties", "underwear", "upskirt", "bottomless", "topless", "petite", "xxx"]


css = """
.btn-green {
  background-image: linear-gradient(to bottom right, #6dd178, #00a613) !important;
  border-color: #22c55e !important;
  color: #166534 !important;
}
.btn-green:hover {
  background-image: linear-gradient(to bottom right, #6dd178, #6dd178) !important;
}
"""

def generate(prompt, samp_steps, batch_size, seed, progress=gr.Progress(track_tqdm=True)):
    prompt = prompt.lower()
    if nsfw_filter:
        if prompt[:10] == "krebzonide":
            prompt = prompt[10:]
        else:
            neg_prompt = neg_prompt + ", child, nsfw, nipples, nude, underwear"
            for word in naughtyWords:
                if prompt.find(word) >= 0:
                    return None, 58008
    if seed < 0:
        seed = random.randint(1,999999)
    images = pipe(
        prompt,
        num_inference_steps=samp_steps,
        num_images_per_prompt=batch_size,
        guidance_scale=0.0,
        generator=torch.manual_seed(seed),
    ).images
    return gr.update(value = [(img, f"Image {i+1}") for i, img in enumerate(images)], height=height+90), seed
        
def set_base_model(base_model_id):
    vae = AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16)
    global model_url_list
    model_url = "https://huggingface.co/" + model_url_list[base_model_id]
    pipe = AutoPipelineForText2Image.from_pretrained(
        "stabilityai/sdxl-turbo",
        torch_dtype = torch.float16,
        variant = "fp16"
        #use_auth_token=hf_token
    )
    pipe.to("cuda")
    return pipe

with gr.Blocks(css=css) as demo:
    with gr.Column():
        prompt = gr.Textbox(label="Prompt")
        submit_btn = gr.Button("Generate", elem_classes="btn-green")
        with gr.Row():
            samp_steps = gr.Slider(1, 30, value=10, step=1, label="Sampling steps")
            batch_size = gr.Slider(1, 6, value=1, step=1, label="Batch size", interactive=True)
            seed = gr.Number(label="Seed", value=-1, minimum=-1, precision=0)
            lastSeed = gr.Number(label="Last Seed", value=-1, interactive=False)
        gallery = gr.Gallery(show_label=False, preview=True, container=False)
        with gr.Row():
    submit_btn.click(generate, [prompt, samp_steps, batch_size, seed], [gallery, lastSeed], queue=True)
    
pipe = set_base_model(model_id)
demo.launch(debug=True)