app.py

import numpy as np
import tensorflow as tf
import gradio as gr
from tensorflow.keras.preprocessing import image
from huggingface_hub import snapshot_download
import os

# Load the model from Hugging Face Hub
def load_model(repo_id):
    download_dir = snapshot_download(repo_id)
    model_path = os.path.join(download_dir, "full_model.weights.h5")
    model = tf.keras.models.load_model(model_path)
    return model

# Function to preprocess the uploaded image
def preprocess_image(img, target_size=(224, 224)):
    img = img.resize(target_size)  # Resize to match model input
    img_array = image.img_to_array(img)
    img_array = np.expand_dims(img_array, axis=0)  # Add batch dimension
    img_array = tf.keras.applications.efficientnet.preprocess_input(img_array)
    return img_array

# Perform inference
def predict(image_input):
    class_names = ["Defective Tyre", "Good Tyre"]
    
    # Preprocess image
    img_array = preprocess_image(image_input)

    # Get prediction
    prediction = model.predict(img_array)[0][0]  # Scalar sigmoid output
    predicted_class_idx = int(prediction >= 0.5)  # 0 if <0.5, 1 if >=0.5
    predicted_class = class_names[predicted_class_idx]  # Get class name

    return f"Predicted Class: {predicted_class} (Confidence: {prediction:.5f})"

# Hugging Face Model Repository ID
REPO_ID = "skngew/9053220B"  # my actual repo ID

# Load the model
model = load_model(REPO_ID)

# Student ID
student_id = "Student ID: 9053220B"

# Markdown description to show classification threshold
threshold_info = """
### EfficientNetB0 (Feature Extraction)
### Classification Threshold:
- A tyre is classified as **Good** if the confidence score is **≥ 0.5**.
- A tyre is classified as **Defective** if the confidence score is **< 0.5**.
"""

# Create the Gradio interface
interface = gr.Interface(
    fn=predict,
    inputs=gr.Image(type="pil", label="Upload an Image"),
    outputs=gr.Textbox(label="Prediction"),
    title="Binary Classification: Good vs. Defective Tyre",
    description=student_id,
    allow_flagging="never",
    examples=[],  #Can add examples here
)

# Add the threshold information markdown
with gr.Blocks() as app:
    gr.Markdown(threshold_info)  # Display threshold info
    interface.render()

# Launch the Gradio app
app.launch(share=True)