Trying to load pre downloaded Imagenet2012 validation set only.

What I need help with / What I was wondering
I am trying to load only validation set for ImageNet2012, and I have already it downloaded and have been able to load it no problem in Pytorch, but I need to load it for TensorFlow models. How do I only load and use the validation set?

What I've tried so far
I have tried tfds builder and load but there doesn't seem to be an option for only loading the validation data set without it trying to download the training data, which I don't have the hard drive space for.

It would be nice if...
I could find documentation or examples of how to load pre downloaded ImageNet 2012 validation set only.

Environment information
(if applicable)

• Operating System: In docker on Ubuntu 18.04 running in WSL 2
• Python version: 3.7 or 3.8 not sure exactly
• tensorflow-datasets/tfds-nightly version: tensorflow-datasets 4.2.0
• tensorflow/tensorflow-gpu/tf-nightly/tf-nightly-gpu version: tf-nightly 2.6.0-dev20210408

Start by reviewing the TensorFlow Datasets builder and load behavior for ImageNet2012, focusing on whether a pre-downloaded validation split can be selected without downloading training data. Check the existing documentation and examples for split selection and local dataset files. Done means adding documentation or an example that shows how to load only the validation set.