app.py

import os
import gradio as gr
import numpy as np
import random
from huggingface_hub import AsyncInferenceClient
from translatepy import Translator
import requests
import re
import asyncio
from PIL import Image
from gradio_client import Client, handle_file
from huggingface_hub import login
from gradio_imageslider import ImageSlider

translator = Translator()
HF_TOKEN = os.environ.get("HF_TOKEN")
HF_TOKEN_UPSCALER = os.environ.get("HF_TOKEN_UPSCALER")
MAX_SEED = np.iinfo(np.int32).max
CSS = "footer { visibility: hidden; }"
JS = "function () { gradioURL = window.location.href; if (!gradioURL.endsWith('?__theme=dark')) { window.location.replace(gradioURL + '?__theme=dark'); } }"

def enable_lora(lora_add, basemodel): 
    return basemodel if not lora_add else lora_add

async def generate_image(prompt, model, lora_word, width, height, scales, steps, seed):
    if seed == -1: 
        seed = random.randint(0, MAX_SEED)
    seed = int(seed)
    text = str(translator.translate(prompt, 'English')) + "," + lora_word
    client = AsyncInferenceClient()
    image = await client.text_to_image(prompt=text, height=height, width=width, guidance_scale=scales, num_inference_steps=steps, model=model)
    return image, seed

async def gen(prompt, basemodel, width, height, scales, steps, seed, upscale_factor, process_upscale, lora_model, process_lora):
    model = enable_lora(lora_model, basemodel) if process_lora else basemodel
    image, seed = await generate_image(prompt, model, "", width, height, scales, steps, seed)
    image_path = "temp_image.jpg"  
    image.save(image_path, format="JPEG")  
    
    if process_upscale:
        upscale_image = get_upscale_finegrain(prompt, image_path, upscale_factor)
    else:
        upscale_image = image_path

    return [image_path, upscale_image]

def get_upscale_finegrain(prompt, img_path, upscale_factor):
    client = Client("finegrain/finegrain-image-enhancer", hf_token=HF_TOKEN_UPSCALER)
    result = client.predict(input_image=handle_file(img_path), prompt=prompt, negative_prompt="", seed=42, upscale_factor=upscale_factor, controlnet_scale=0.6, controlnet_decay=1, condition_scale=6, tile_width=112, tile_height=144, denoise_strength=0.35, num_inference_steps=18, solver="DDIM", api_name="/process")
    return result[1]

css = """
#col-container{
    margin: 0 auto;
    max-width: 1024px;
}
"""

with gr.Blocks(css=CSS, js=JS, theme="Nymbo/Nymbo_Theme") as demo:
    with gr.Column(elem_id="col-container"):
        gr.Markdown("Flux Upscaled +LORA")
        with gr.Row():
            with gr.Column(scale=1.5):
                output_res = ImageSlider(label="Flux / Upscaled")
            with gr.Column(scale=0.8):
                prompt = gr.Textbox(label="Prompt")
                basemodel_choice = gr.Dropdown(label="Base Model", choices=["black-forest-labs/FLUX.1-schnell", "black-forest-labs/FLUX.1-DEV"], value="black-forest-labs/FLUX.1-schnell")
                lora_model_choice = gr.Dropdown(label="LORA Model", choices=["Shakker-Labs/FLUX.1-dev-LoRA-add-details", "XLabs-AI/flux-RealismLora"], value="XLabs-AI/flux-RealismLora")
                process_lora = gr.Checkbox(label="Process LORA")
                process_upscale = gr.Checkbox(label="Process Upscale")
                upscale_factor = gr.Radio(label="UpScale Factor", choices=[2, 4, 8], value=2)
                
                with gr.Accordion(label="Advanced Options", open=False):
                    width = gr.Slider(label="Width", minimum=512, maximum=1280, step=8, value=1280)
                    height = gr.Slider(label="Height", minimum=512, maximum=1280, step=8, value=768)
                    scales = gr.Slider(label="Guidance", minimum=3.5, maximum=7, step=0.1, value=3.5)
                    steps = gr.Slider(label="Steps", minimum=1, maximum=100, step=1, value=24)
                    seed = gr.Slider(label="Seeds", minimum=-1, maximum=MAX_SEED, step=1, value=-1)

                submit_btn = gr.Button("Submit", scale=1)
                submit_btn.click(
                    fn=lambda: None,
                    inputs=None,
                    outputs=[output_res],
                    queue=False
                ).then(
                    fn=gen,
                    inputs=[prompt, basemodel_choice, width, height, scales, steps, seed, upscale_factor, process_upscale, lora_model_choice, process_lora],
                    outputs=[output_res]
                )