Load local dataset on TensorFlow/MoviNet

Hi all!

I’m new at deep learning, and I’m facing some issues with that.

I’m working with transfer learning using MoviNet. It happens that I’ve got a large dataset on my on drive and I wanted it to be loaded and used by MoviNet.

I’ve googled A LOT but found less than nothing. It seems that the only way of invoke my dataset is adding it to TensorFlow Datasets.

I wish to use functions such as builder and download _and_prepare. But, for this, I’d have to make a lot o tricky things.

Is there a way of simply call my own dataset providing its path, such as “C://path_of_data” or “https://link_of_data”?

I really appreciate any kind of suggestions. Please help me :smiling_face_with_tear:

Hi @Felipe-m,

Please use this Tutorial to convert your data into tf records and load as your tf data using following link

Please let me know if this is helpful.