app.py

from flask import Flask, request, jsonify, send_file
from PIL import Image
import base64
import io
import random
import uuid
import numpy as np
import spaces
import torch
from diffusers import StableDiffusionXLPipeline, EulerAncestralDiscreteScheduler

# Create a Flask instance
app = Flask(__name__)

def clear_gpu_memory():
    torch.cuda.empty_cache()
    torch.cuda.synchronize()

# Initialize model only once
pipe = None

def load_model():
    global pipe
    if pipe is None:
        pipe = StableDiffusionXLPipeline.from_pretrained(
            "fluently/Fluently-XL-v2",
            torch_dtype=torch.float16,
            use_safetensors=True,
        )
        pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config)
        pipe.load_lora_weights("ehristoforu/dalle-3-xl-v2", weight_name="dalle-3-xl-lora-v2.safetensors", adapter_name="dalle")
        pipe.set_adapters("dalle")

        if torch.cuda.is_available():
            pipe.to("cuda")

# Load the model during app initialization
load_model()

def save_image(img):
    unique_name = str(uuid.uuid4()) + ".png"
    img.save(unique_name)
    return unique_name

def randomize_seed_fn(seed: int, randomize_seed: bool) -> int:
    if randomize_seed:
        seed = random.randint(0, MAX_SEED)
    return seed

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

@spaces.GPU
def generate(
    prompt: str,
    negative_prompt: str = "",
    use_negative_prompt: bool = False,
    seed: int = 0,
    num_images_per_prompt: int = 1,
    width: int = 512,  # Reduced image width
    height: int = 512,  # Reduced image height
    guidance_scale: float = 3,
    randomize_seed: bool = False,
):
    seed = int(randomize_seed_fn(seed, randomize_seed))

    if not use_negative_prompt:
        negative_prompt = ""  # type: ignore

    images = pipe(
        prompt=prompt,
        negative_prompt=negative_prompt,
        width=width,
        height=height,
        guidance_scale=guidance_scale,
        num_inference_steps=25,
        num_images_per_prompt=num_images_per_prompt,
        cross_attention_kwargs={"scale": 0.65},
        output_type="pil",
    ).images
    image_paths = [save_image(img) for img in images]
    print(image_paths)
    return image_paths, seed

@app.get("/")
def root():
    return "Welcome to the Fashion Outfit"

@app.route('/api/get_image/<image_id>', methods=['GET'])
def get_image(image_id):
    try:
        return send_file(image_id, mimetype='image/png')
    except FileNotFoundError:
        return jsonify({'error': 'Image not found'}), 404

@app.route('/api/run', methods=['POST'])
def run():
    data = request.json
    print(data)
    prompt = data['prompt']
    negative_prompt = data['negative_prompt']
    use_negative_prompt = data['use_negative_prompt']
    guidance_scale = data['guidance_scale']
    randomize_seed = data['randomize_seed']
    num_images_per_prompt = data['num_images_per_prompt']
    width = data['width'] if 'width' in data else 512  # Default width
    height = data['height'] if 'height' in data else 512  # Default height
    #clear_gpu_memory()
    result = generate(
        prompt,
        negative_prompt,
        use_negative_prompt,
        0,
        num_images_per_prompt,
        width,
        height,
        guidance_scale,
        randomize_seed
    )
    return jsonify({'out': result})

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=7860)