app.py

import os
import io
import random
import requests
import gradio as gr
import numpy as np
from PIL import Image
import replicate


MAX_SEED = np.iinfo(np.int32).max


def predict(replicate_api, prompt, lora_id, lora_scale=0.95, aspect_ratio="1:1", seed=-1, randomize_seed=True, guidance_scale=3.5, num_inference_steps=28, progress=gr.Progress(track_tqdm=True)):

    # Validate API key and prompt
    if not replicate_api or not prompt:
        return "Error: Missing necessary inputs.", -1, None
    
    # Set the seed if randomize_seed is True
    if randomize_seed:
        seed = random.randint(0, MAX_SEED)

    # Set the Replicate API token in the environment variable
    os.environ["REPLICATE_API_TOKEN"] = replicate_api

    # Construct the input for the replicate model
    input_params = {
        "prompt": prompt,
        "output_format": "jpg",
        "aspect_ratio": aspect_ratio,
        "num_inference_steps": num_inference_steps,
        "guidance_scale": guidance_scale,
        "seed": seed,
        "disable_safety_checker": True
    }

    # If lora_id is provided, include it in the input
    if lora_id and lora_id.strip()!="":
        input_params["hf_lora"] = lora_id.strip()
        input_params["lora_scale"] = lora_scale

    try:
        # Run the model using the user's API token from the environment variable
        output = replicate.run(
            "lucataco/flux-dev-lora:a22c463f11808638ad5e2ebd582e07a469031f48dd567366fb4c6fdab91d614d",
            input=input_params
        )
        print(output,prompt)
        return output[0], seed, seed  # Return the generated image and seed

    except Exception as e:
        # Catch any exceptions, such as invalid API token or lack of credits
        return f"Error: {str(e)}", -1, None

    finally:
        # Always remove the API key from the environment
        if "REPLICATE_API_TOKEN" in os.environ:
            del os.environ["REPLICATE_API_TOKEN"]

    

demo = gr.Interface(fn=predict, inputs="text", outputs="image")

css="""
#col-container {
    margin: 0 auto;
    max-width: 520px;
}
"""

examples = [
    "a tiny astronaut hatching from an egg on the moon",
    "a cat holding a sign that says hello world",
    "an anime illustration of a wiener schnitzel",
]

with gr.Blocks(css=css) as demo:
    with gr.Column(elem_id="col-container"):
        gr.Markdown("# FLUX Dev with Replicate API")
        
        replicate_api = gr.Text(label="Replicate API Key", type='password', show_label=True, max_lines=1, placeholder="Enter your Replicate API token", container=True)
        prompt = gr.Text(label="Prompt", show_label=True, lines = 2, max_lines=4, show_copy_button = True, placeholder="Enter your prompt", container=True)
        with gr.Accordion("Advanced Settings", open=False):
            with gr.Row():
                custom_lora = gr.Textbox(label="Custom LoRA", info="LoRA Hugging Face path (optional)", placeholder="multimodalart/vintage-ads-flux")
                lora_scale = gr.Slider(
                    label="LoRA Scale",
                    minimum=0,
                    maximum=1,
                    step=0.01,
                    value=0.95,
                )
            aspect_ratio = gr.Radio(label="Aspect ratio", value="1:1", choices=["1:1", "4:5", "2:3", "3:4","9:16", "4:3", "16:9"])
            seed = gr.Slider(
                label="Seed",
                minimum=0,
                maximum=MAX_SEED,
                step=1,
                value=0,
            )

            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)

            with gr.Row():
                guidance_scale = gr.Slider(
                    label="Guidance Scale",
                    minimum=1,
                    maximum=15,
                    step=0.1,
                    value=3.5,
                )
                num_inference_steps = gr.Slider(
                    label="Number of inference steps",
                    minimum=1,
                    maximum=50,
                    step=1,
                    value=28,
                )
        submit = gr.Button("Generate Image", variant="primary",scale=1)

        output = gr.Image(label="Output Image", show_label=True)

        seed_used = gr.Textbox(label="Seed Used", show_copy_button = True)
        

        gr.Examples(
            examples=examples,
            fn=predict,
            inputs=[prompt]
        )
        gr.on(
            triggers=[submit.click, prompt.submit],
            fn=predict,
            inputs=[replicate_api, prompt, custom_lora, lora_scale, aspect_ratio, seed, randomize_seed, guidance_scale, num_inference_steps],
            outputs = [output, seed, seed_used]
        )

demo.launch()