app.py

import gradio as gr
import tensorflow as tf
import numpy as np
from PIL import Image

model_path = "p_inference_fruits/Xeption_fruits.keras"
model = tf.keras.models.load_model(model_path)

# Define the core prediction function
def predict_fruit(image):
    # Preprocess image
    print(type(image))
    image = Image.fromarray(image.astype('uint8'))  # Convert numpy array to PIL image
    image = image.resize((150, 150)) #resize the image to 28x28 and converts it to gray scale
    image = np.array(image)
    image = np.expand_dims(image, axis=0) # same as image[None, ...]
    
    # Predict
    prediction = model.predict(image)
    
     # No need to apply sigmoid, as the output layer already uses softmax
    # Convert the probabilities to rounded values
    prediction = np.round(prediction, 3)

    # Separate the probabilities for each class
    p_apple = prediction[0][0]  # Probability for class 'articuno'
    p_banana = prediction[0][1]   # Probability for class 'moltres'
    p_pinenapple = prediction[0][2]    # Probability for class 'zapdos'
    p_strawberries = prediction[0][3]
    p_watermelon = prediction[0][4]
 
    return {'apple':  p_apple, 'banana': p_banana, 'pinenapple': p_pinenapple, 'strawberries': p_strawberries, 'watermelon': p_watermelon}

# Create the Gradio interface
input_image = gr.Image()
iface = gr.Interface(
    fn=predict_fruit,
    inputs=input_image,
    outputs=gr.Label(),
    examples=["p_inference_fruits/images/ap1.jpeg", "p_inference_fruits/images/ap2.jpeg", "p_inference_fruits/images/ap3.jpeg", "p_inference_fruits/images/ba1.jpeg", "p_inference_fruits/images/ba2.jpeg", "p_inference_fruits/images/ba3.jpeg", "p_inference_fruits/images/pi1.jpeg","p_inference_fruits/images/pi2.jpeg","p_inference_fruits/images/pi3.jpeg","p_inference_fruits/images/st1.jpeg", "p_inference_fruits/images/st2.jpeg", "p_inference_fruits/images/st3.jpeg","p_inference_fruits/images/wa1.jpeg","p_inference_fruits/images/wa2.jpeg","p_inference_fruits/images/wa3.jpeg"],  
    description="TEST.")

iface.launch()