app.py

import argparse
import os
import random
import torch
import numpy as np
from diffusers import DiffusionPipeline, AutoencoderKL
from PIL import Image
import re

def generate_image(pipe, prompt, seed=42, randomize_seed=True, width=768, height=768, guidance_scale=4.5, num_inference_steps=20):
    MAX_SEED = np.iinfo(np.int32).max
    
    # 无论如何都先创建 generator
    generator = torch.Generator(device=pipe.device)

    if randomize_seed:
        seed = random.randint(0, MAX_SEED)
    
    # 使用最终确定的种子来设置 generator
    generator.manual_seed(seed)
    print(f"ℹ️  使用种子: {seed}")
    
    print("🚀 开始生成图像...")
    image = pipe(
        prompt=prompt,
        guidance_scale=guidance_scale,
        num_inference_steps=num_inference_steps,
        width=width,
        height=height,
        generator=generator,
        output_type="pil"
    ).images[0]
    
    return image, seed

def main():
    parser = argparse.ArgumentParser(description="使用 FLUX.1-Krea-dev 模型从文本提示生成图像。")
    parser.add_argument("--prompt", type=str, required=True, help="用于图像生成的文本提示。")
    parser.add_argument("--seed", type=int, default=None, help="随机种子。如果未提供，将随机生成。")
    parser.add_argument("--steps", type=int, default=20, help="推理步数。")
    parser.add_argument("--width", type=int, default=768, help="图像宽度。")
    parser.add_argument("--height", type=int, default=768, help="图像高度。")
    parser.add_argument("--guidance", type=float, default=4.5, help="指导比例 (Guidance Scale)。")
    args = parser.parse_args()

    print("⏳ 正在加载模型，请稍候...")
    dtype = torch.bfloat16
    device = "cuda" if torch.cuda.is_available() else "cpu"
    
    good_vae = AutoencoderKL.from_pretrained("black-forest-labs/FLUX.1-Krea-dev", subfolder="vae", torch_dtype=dtype)
    pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-Krea-dev", torch_dtype=dtype, vae=good_vae).to(device)

    if device == "cuda":
        torch.cuda.empty_cache()

    print(f"✅ 模型加载完成，使用设备: {device}")

    print(f"🎨 开始为提示生成图像: '{args.prompt}'")

    randomize = args.seed is None
    seed_value = args.seed if not randomize else 42 

    generated_image, used_seed = generate_image(
        pipe=pipe,
        prompt=args.prompt,
        seed=seed_value,
        randomize_seed=randomize,
        width=args.width,
        height=args.height,
        num_inference_steps=args.steps,
        guidance_scale=args.guidance
    )

    output_dir = "output"
    os.makedirs(output_dir, exist_ok=True)
    
    safe_prompt = re.sub(r'[^\w\s-]', '', args.prompt).strip()
    safe_prompt = re.sub(r'[-\s]+', '_', safe_prompt)
    
    filename = f"{safe_prompt[:50]}_{used_seed}.png"
    filepath = os.path.join(output_dir, filename)

    print(f"💾 正在保存图像到: {filepath}")
    generated_image.save(filepath)

    print("🎉 完成！")

if __name__ == "__main__":
    main()