How should tfrecord data be processed or read for the keras multi input model

General Discussion
, , , ,
#1

image
image896×199 14 KB

#3

Hi @11127

Welcome to the TensorFlow Forum!

You need to use tf.io.parse_example() method to decode the TFRecord Data into a dictionary of standard tensors, split the dataset into features for multi-input model and then feed the into model.fit for the training.

Please refer to this Reading a TFRecord file Using tf.data doc for better understanding. Thank you.