evaluate.py

from collections import Counter
import json
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import pickle
import tensorflow as tf
import tensorflow_datasets as tfds
import string

MAX_LENGTH = 80
def loss_function(y_true, y_pred):
    y_true = tf.reshape(y_true, shape=(-1, MAX_LENGTH - 1))

    loss = tf.keras.losses.SparseCategoricalCrossentropy(
        from_logits=True, reduction='none')(y_true, y_pred)

    mask = tf.cast(tf.not_equal(y_true, 0), tf.float32)
    loss = tf.multiply(loss, mask)

    return tf.reduce_mean(loss)

class CustomSchedule(tf.keras.optimizers.schedules.LearningRateSchedule):

    def __init__(self, d_model, warmup_steps=4000):
        super(CustomSchedule, self).__init__()

        self.d_model = tf.constant(d_model, dtype=tf.float32)
        self.warmup_steps = warmup_steps

    def get_config(self):
        return {"d_model": self.d_model, "warmup_steps": self.warmup_steps}

    def __call__(self, step):
        arg1 = tf.math.rsqrt(step)
        arg2 = step * (self.warmup_steps ** -1.5)

        return tf.math.multiply(tf.math.rsqrt(self.d_model), tf.math.minimum(arg1, arg2))

def accuracy(y_true, y_pred):
    y_true = tf.reshape(y_true, shape=(-1, MAX_LENGTH - 1))
    return tf.keras.metrics.sparse_categorical_accuracy(y_true, y_pred)

def evaluate(sentence):
    sentence = preprocess_sentence(sentence)

    sentence = tf.expand_dims(
        START_TOKEN + tokenizer.encode(sentence) + END_TOKEN, axis=0)

    output = tf.expand_dims(START_TOKEN, 0)

    for i in range(MAX_LENGTH):
        predictions = model(inputs=[sentence, output], training=False)

        predictions = predictions[:, -1:, :]
        predicted_id = tf.cast(tf.argmax(predictions, axis=-1), tf.int32)

        if tf.equal(predicted_id, END_TOKEN[0]):
            break

        output = tf.concat([output, predicted_id], axis=-1)

def predict(sentence):
    predict = evaluate(sentence)
    predicted_sentence = tokenizer.decode([i for i in predict if i < tokenizer.vocab_size])
    return print('Bot: {}'.format(predicted_sentence.replace('\&undsc', ' ')))