import torch
import torch.nn as nn
from torchvision import models, transforms
from torch.utils.data import DataLoader
from PIL import Image
import gradio as gr

DEVICE = "cuda" if torch.cuda.is_available() else "cpu"

# Load class names (make sure this file is in the Space)
with open("cifar10_classes.txt") as f:
    CLASSES = [line.strip() for line in f.readlines()]

def build_model(num_classes: int, device: str = "cpu"):
    try:
        weights = models.ResNet18_Weights.DEFAULT
        model = models.resnet18(weights=weights)
    except AttributeError:
        model = models.resnet18(weights="IMAGENET1K_V1")
    model.fc = nn.Linear(model.fc.in_features, num_classes)
    model = model.to(device)
    return model

num_classes = len(CLASSES)
model = build_model(num_classes, device=DEVICE)

state_dict = torch.load("ast_cifar10_resnet18.pth", map_location=DEVICE)
model.load_state_dict(state_dict)
model.eval()

preprocess = transforms.Compose([
    transforms.Resize((224, 224)),
    transforms.ToTensor(),
])

def predict(image: Image.Image):
    if image is None:
        return {}
    x = preprocess(image).unsqueeze(0).to(DEVICE)
    with torch.no_grad():
        logits = model(x)
        probs = torch.softmax(logits, dim=1)[0]
    return {CLASSES[i]: float(probs[i]) for i in range(len(CLASSES))}

demo = gr.Interface(
    fn=predict,
    inputs=gr.Image(type="pil", label="Upload CIFAR-like image"),
    outputs=gr.Label(num_top_classes=3, label="Top-3 Predictions"),
    title="AST CIFAR-10 Classifier",
    description="ResNet18 fine-tuned with Adaptive Sparse Training (AST) on CIFAR-10.",
)

if __name__ == "__main__":
    demo.launch()