app.py

### 1. Imports and class names setup ### 
import gradio as gr
import os
import torch
import numpy as np
from PIL import Image
from model import create_vit_model
from timeit import default_timer as timer
from typing import Tuple, Dict

# Setup class names
with open("class_names.txt", "r") as f:
    class_names = [food_name.strip() for food_name in f.readlines()]
    
### 2. Model and transforms preparation ###    

# Create model
vit, vit_transforms = create_vit_model(num_classes=121)

# Load saved weights
vit.load_state_dict(
    torch.load(
        f="vit_epoch_2.pth",
        map_location=torch.device("cpu"),
    )
)

### 3. Predict function ###

from PIL import Image
import numpy as np

def predict(img) -> Tuple[Dict, float]:
    """Transforms and performs a prediction on img and returns prediction and time taken."""
    start_time = timer()

    if isinstance(img, np.ndarray):
        img = img.astype(np.uint8)  # ✅ Ensure dtype is uint8
        img = Image.fromarray(img, mode="RGB")  # ✅ Safe conversion

    img = vit_transforms(img).unsqueeze(0)

    vit.eval()
    with torch.inference_mode():
        pred_probs = torch.softmax(vit(img), dim=1)

    pred_labels_and_probs = {
        class_names[i]: float(pred_probs[0][i]) for i in range(len(class_names))
    }

    pred_time = round(timer() - start_time, 5)
    return pred_labels_and_probs, pred_time


### 4. Gradio app ###

title = "VisionBite 🍔👁"
description = "A ViT feature extractor computer vision model to classify images of food into 121 categories."
article = "The model has been trained on the Food121 dataset using ViT Base 16."

# ✅ Sort examples for consistent UI (optional)
example_list = [["examples/" + example] for example in sorted(os.listdir("examples")) if example.endswith((".jpg", ".png", ".jpeg"))]

demo = gr.Interface(
    fn=predict,
    inputs=gr.Image(type="pil"),
    outputs=[
        gr.Label(num_top_classes=5, label="Predictions"),
        gr.Number(label="Prediction time (s)"),
    ],
    examples=example_list,
    title=title,
    description=description,
    article=article,
)

demo.launch()