app.py

# -*- coding: utf-8 -*-
"""
Spyder Editor

This is a temporary script file.
"""

import gradio as gr
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers
from tensorflow.keras.layers import TextVectorization, Embedding, Dense
from custom_classes import TransformerEncoder, PositionalEmbedding

model = keras.models.load_model(
 "full_transformer_encoder.keras",
 custom_objects={"TransformerEncoder": TransformerEncoder,
 "PositionalEmbedding": PositionalEmbedding})

def make_prediction(input_text):
 myTensor = tf.convert_to_tensor(input_text, dtype=tf.string)
 pred = model(tf.reshape(myTensor, (-1,1)))
 label_index = int(pred.numpy()[0,0] + 0.5)
 mapping = {0: 'Negative', 1: 'Positive'}
 label = mapping[label_index]
 return label


#Create the Gradio demo
demo = gr.Interface(fn=make_prediction,
 inputs="text",
 outputs="text",
 title="Text Classification",
 description="built via gradio")
demo.launch()