app.py

from PIL import Image
import matplotlib.pyplot as plt
import gradio as gr
from transformers import pipeline
from transformers import AutoModelForImageClassification, AutoImageProcessor

image_processor = AutoImageProcessor.from_pretrained("./Mymodel/")
model = AutoModelForImageClassification.from_pretrained("./Mymodel/")

def predict(my_image):

    image = Image.fromarray(image.astype('uint8'))

    pipe = pipeline("image-classification", 
                    model=model,
                    feature_extractor=image_processor)
    
    pred = pipe(image)
    
    plt.imshow(image)
    plt.title(pred[0]['label'].replace('_', ' ').title())
    plt.axis(False)
    plt.show()

    print(f"Predicted the above image as a {pred[0]['label'].replace('_', ' ').title()} with {pred[0]['score']*100:.2f}% confidence")
    
    run = True
    while run:

        inp = input('Is the prediction correct?')
        if inp.lower() == 'yes':
            print(f"""
{food_info[pred[0]['label'].replace('_', ' ').title()]['Description']}

Info: {food_info[pred[0]['label'].replace('_', ' ').title()]['Calories and Health Info']}""")
            run = False
        elif inp.lower() == 'no':
            print(f"""
The image could be a {pred[1]['label'].replace('_', ' ').title()}, with a {pred[1]['score']*100:.2f}% confidence,
The image could be a {pred[2]['label'].replace('_', ' ').title()}, with a {pred[2]['score']*100:.2f}% confidence,
The image could be a {pred[3]['label'].replace('_', ' ').title()}, with a {pred[3]['score']*100:.2f}% confidence,
Or the image could be a {pred[4]['label'].replace('_', ' ').title()}, with a {pred[4]['score']*100:.2f}% confidence,
                   """)
            run = False
        else:
            print('Please respond as yes or no')

iface = gr.Interface(fn=predict, inputs="image", outputs="image")
iface.launch()