app.py

# -*- coding: utf-8 -*-
"""app.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1Mv-09djxvNrRlLGT7tkVUiy9YGyOLW5G
"""

import gradio as gr
import cv2
import requests
import gdown
import tensorflow as tf
from tensorflow import keras
#from custom_model import ImageClassifier
import numpy as np
#from tensorflow.keras import models
#from keras.layers import Dense, Activation, Flatten,Dropout, Conv2D, BatchNormalization, MaxPooling2D

from keras.models import load_model

# Add any necessary imports
import tensorflow as tf
import gradio as gr
from tensorflow.keras.models import load_model
import numpy as np
import tensorflow_hub as hub

#path  = [['car_bike.jpg'], ['human.jpg'], ['chair.jpg']]

# Define your class labels or categories for predictions
train_info = []  # Replace with your actual class labels

# Read image names from the text file
with open('label.txt', 'r') as file:
    train_info = [line.strip() for line in file.read().splitlines()]

output_path = 'label.txt'
with open(output_path, 'r') as file:
    LABELS = [x.strip() for x in file.readlines()]

num_classes = 12
IMG_SIZE = 224

def _normalize_img(img):
    img = tf.cast(img, tf.float32) / 255.  # All images will be rescaled by 1./255
    img = tf.image.resize(img, (IMG_SIZE, IMG_SIZE), method='bilinear')
    return img

# If your model requires custom objects, provide them here
model = load_model('model.h5', custom_objects={'KerasLayer': hub.KerasLayer})

def predict_fn(img):
    img = img.convert('RGB')
    img_data = _normalize_img(img)
    x = np.array(img_data)
    x = np.expand_dims(x, axis=0)
    temp = model.predict(x)

    idx = np.argsort(np.squeeze(temp))[::-1]
    top3_value = np.asarray([temp[0][i] for i in idx[0:3]])
    top3_idx = idx[0:3]

    return {LABELS[i]: str(v) for i, v in zip(top3_idx, top3_value)}

# Launch the Gradio interface
gr.Interface(predict_fn, gr.inputs.Image(type='pil'), outputs='label').launch()