app.py

import streamlit as st
import torch
import numpy as np
from PIL import Image
from model import CycleGAN, get_val_transform, de_normalize


st.set_page_config(
    page_title="CycleGAN Image Converter",
    page_icon="🎨",
    layout="wide"
)


@st.cache_resource
def get_device():
    if torch.cuda.is_available():
        device = torch.device("cuda")
        st.sidebar.success("Using GPU 🚀")
    elif torch.backends.mps.is_available():
        device = torch.device("mps")
        st.sidebar.success("Using Apple Silicon 🍎")
    else:
        device = torch.device("cpu")
        st.sidebar.info("Using CPU 💻")
    return device


st.markdown("""
    <style>
    .stApp {
        max-width: 1200px;
        margin: 0 auto;
    }
    .main {
        padding: 2rem;
    }
    </style>
""", unsafe_allow_html=True)


st.title("CycleGAN Image Converter 🎨")
st.markdown("""
    Transform images between different domains using CycleGAN.
    Upload an image and see it converted in real-time!
    
    *Note: Images will be resized to 256x256 pixels during conversion.*
""")


MODELS = [
    {
        "name": "Cezanne ↔ Photo",
        "id": "cezanne2photo",
        "model_path": "waleko/cyclegan",
        "description": "Convert between Cezanne's painting style and photographs"
    },
    {
        "name": "Day ↔ Night",
        "id": "day2night",
        "model_path": "waleko/cyclegan-day_night",
        "description": "Convert between day and night cityscapes"
    }
]


with st.sidebar:
    st.header("Settings")
    
    
    selected_model = st.selectbox(
        "Conversion Type",
        options=range(len(MODELS)),
        format_func=lambda x: MODELS[x]["name"]
    )
    
    
    direction = st.radio(
        "Conversion Direction",
        options=["A → B", "B → A"],
        help="A → B: Convert from domain A to B\nB → A: Convert from domain B to A"
    )


@st.cache_resource
def load_model(model_path):
    device = get_device()
    model = CycleGAN.from_pretrained(model_path)
    model = model.to(device)
    model.eval()
    return model


def process_image(image, model, direction):
    transform = get_val_transform(model, direction)
    tensor = transform(np.array(image)).unsqueeze(0)
    tensor = tensor.to(next(model.parameters()).device)
    with torch.no_grad():
        if direction == "A → B":
            output = model.generator_ab(tensor)
        else:
            output = model.generator_ba(tensor)
    result = de_normalize(output[0], model, direction)
    image = Image.fromarray(result.cpu().detach().numpy(), 'RGB')
    return image


col1, col2 = st.columns(2)

with col1:
    st.subheader("Input Image")
    uploaded_file = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])
    
    if uploaded_file is not None:
        input_image = Image.open(uploaded_file)
        
        if input_image.mode == 'RGBA':
            input_image = input_image.convert('RGB')
        st.image(input_image)

with col2:
    st.subheader("Converted Image")
    if uploaded_file is not None:
        try:
            
            model = load_model(MODELS[selected_model]["model_path"])
            result = process_image(input_image, model, direction)
            
            
            st.image(result)
        except Exception as e:
            st.error(f"Error during conversion: {str(e)} {e.__traceback__}")
            raise
    else:
        st.info("Upload an image to see the conversion result")