| import gradio as gr | |
| from tensorflow import keras as k | |
| import numpy as np | |
| loaded_CNN = k.models.load_model('Digit_classification_model2.h5') | |
| def predict(img): | |
| img_array = np.array(img) | |
| img_array = img_array.reshape(1, 28, 28) | |
| img_array = img_array/255 | |
| pred = loaded_CNN.predict(img_array) | |
| print(pred) | |
| return np.argmax(pred) | |
| iface = gr.Interface(predict, inputs = 'sketchpad', | |
| outputs = 'text', | |
| allow_flagging = 'never', | |
| description = 'Project : Recognizing hardwritten digits : Draw a Single Digit Below... (Draw in the centre for best results)') | |
| iface.launch(debug = "True", width = 500, height = 500) | |