tfrecords_datasets_split

These scrips help users split the trainning and testing(or called validation) datasets in TensorFlow or Keras

---

TensorFlow

• In TensorFlow, one common way to efficiently use data and datasets is to stored them in TFRecord files.
• When used, I access them throught batches and send them to our train or test model.

Keras

• In Keras, it support more simply ways to generate our datasets.
• Typcially, here I use the flow_from_directory method to prepare training and validation datasets.

About the path

• Here I specify the path I used to generate TFRecord files and Keras's training and validation datasets

  /user

  /image_data

  /dog
  /cat
  /elephant
  ...
  /tiger