TPU Speed Data Pipelines Reviews

Nikolas N. · Reviewed almost 2 years ago

Instructions again make absolutely no sense - "iterate the files" is not an instruction that is complete is any way. Even the question numbers are wrong just listing question 1 repeatedly.

Oliver H. · Reviewed almost 2 years ago

aishwarya b. · Reviewed almost 2 years ago

Joost B. · Reviewed almost 2 years ago

Thorvald Sahl J. · Reviewed almost 2 years ago

It's kind of messy. We're not using accelerators (GPUs) and we're not using parallelism at all. I guess it's nice to learn the data loading.

Loading A. · Reviewed almost 2 years ago

Aditya S. · Reviewed almost 2 years ago

Brody I. · Reviewed almost 2 years ago

Decent but there's a lot missing. TFRecord is introduced and then there is no explanation of tf.io.*, so you'll end up searching through a bunch of docs until maybe you come across parse_single_example or until you look at the solution.

chris w. · Reviewed almost 2 years ago

LUCAS S. · Reviewed almost 2 years ago

Sudharshan N. · Reviewed almost 2 years ago

Alexey T. · Reviewed almost 2 years ago

Suchita B. · Reviewed almost 2 years ago

DANIEL M. · Reviewed almost 2 years ago

Jayesh Arun B. · Reviewed almost 2 years ago

ROBERTO O. · Reviewed almost 2 years ago

Rahul M. · Reviewed almost 2 years ago

David M. · Reviewed almost 2 years ago

Harold H. · Reviewed almost 2 years ago

2. To use TFRecord format to load training data efficiently from GCS it is more efficient is it because that it doesn't show the pictures, but it shows 000100 010000 01010?

Wanda W. · Reviewed almost 2 years ago

Ming H. · Reviewed almost 2 years ago

LUIS F. · Reviewed almost 2 years ago

Kidus B. · Reviewed almost 2 years ago

Menegozzo G. · Reviewed almost 2 years ago

Marián Ž. · Reviewed almost 2 years ago