losses.py

import tensorflow as tf
import keras


class CTCloss(tf.keras.losses.Loss):
    """ CTCLoss objec for training the model"""
    def __init__(self, name: str = "CTCloss",reduction=tf.keras.losses.Reduction.SUM) -> None:
        super(CTCloss, self).__init__()
        self.name = name
        self.loss_fn = tf.keras.backend.ctc_batch_cost
        self.reduction = reduction

    def __call__(self, y_true: tf.Tensor, y_pred: tf.Tensor, sample_weight=None) -> tf.Tensor:
        """ Compute the training batch CTC loss value"""
        batch_len = tf.cast(tf.shape(y_true)[0], dtype="int64")
        input_length = tf.cast(tf.shape(y_pred)[1], dtype="int64")
        label_length = tf.cast(tf.shape(y_true)[1], dtype="int64")

        input_length = input_length * tf.ones(shape=(batch_len, 1), dtype="int64")
        label_length = label_length * tf.ones(shape=(batch_len, 1), dtype="int64")

        #print("input_length:",input_length)
        #tf.print(input_length)

        #print("label_length:",label_length)
        #tf.print(label_length)

        loss = self.loss_fn(y_true, y_pred, input_length, label_length)

        return loss