app.py

import os
import torch
import torch.nn as nn
from torchvision import transforms
from PIL import Image
import gradio as gr

# -------- CONFIG --------
checkpoint_path = "age_prediction_model2.pth"  # Just the model file name for Hugging Face Spaces
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print(f"Using device: {device}")

# -------- SIMPLE CNN MODEL --------
class SimpleCNN(nn.Module):
    def __init__(self):
        super(SimpleCNN, self).__init__()
        self.features = nn.Sequential(
            nn.Conv2d(3, 32, kernel_size=3, padding=1), nn.ReLU(),
            nn.MaxPool2d(2),  # 64x64
            nn.Conv2d(32, 64, kernel_size=3, padding=1), nn.ReLU(),
            nn.MaxPool2d(2),  # 32x32
            nn.Conv2d(64, 128, kernel_size=3, padding=1), nn.ReLU(),
            nn.MaxPool2d(2),  # 16x16
        )
        self.classifier = nn.Sequential(
            nn.Flatten(),
            nn.Linear(128 * 16 * 16, 256), nn.ReLU(),
            nn.Linear(256, 1)  # Output: age (regression)
        )

    def forward(self, x):
        x = self.features(x)
        x = self.classifier(x)
        return x

# -------- LOAD MODEL --------
model = SimpleCNN().to(device)

# Check if the checkpoint file exists and load
if os.path.exists(checkpoint_path):
    model.load_state_dict(torch.load(checkpoint_path, map_location=device))  # Load to the correct device
    model.eval()  # Set the model to evaluation mode
    print(f"Model loaded from {checkpoint_path}")
else:
    print(f"Error: Checkpoint file not found at {checkpoint_path}. Please check the path.")

# -------- PREPROCESSING --------
transform = transforms.Compose([
    transforms.Resize((128, 128)),
    transforms.ToTensor(),
    transforms.Normalize([0.5, 0.5, 0.5], [0.5, 0.5, 0.5])
])

def predict_age(image: Image.Image) -> float:
    image_tensor = transform(image).unsqueeze(0).to(device)
    with torch.no_grad():
        output = model(image_tensor)
        age = output.item()
    return round(age, 2)


import gradio as gr

gr.Interface(
    fn=predict_age,
    inputs=gr.Image(type="pil", image_mode="RGB"),
    outputs="number",  # or "text" if your output is text-based
    title="Age Prediction from Face",
    description="Upload a face image and get the predicted age."
).launch()