Creating Input Pipeline using TF-Dataset

**Using python libraries you have to make a pipeline to process the images to feed into the model**

Following are the things you have to do for making this pipeline

- List all image filenames
- Shuffle
- Split into Train/Test
- Create TensorFlow Datasets from lists
- Map 'load' function to TensorFlow Datasets
- Batch Datasets

Start by locating the existing image-loading code and the model input entry point; the issue specifically calls for a `load` function mapped over TensorFlow Datasets. Done means filenames are listed, shuffled, split into train and test sets, converted to TensorFlow Datasets, mapped through `load`, and batched.

Written by the indexing model from the issue text.