关于“TPU Speed Data Pipelines”的评价

Mikael T. · 已于 about 2 years前审核

Ehsan u. · 已于 about 2 years前审核

Nikolas N. · 已于 about 2 years前审核

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. · 已于 about 2 years前审核

aishwarya b. · 已于 about 2 years前审核

Joost B. · 已于 about 2 years前审核

Thorvald Sahl J. · 已于 about 2 years前审核

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. · 已于 about 2 years前审核

Aditya S. · 已于 over 2 years前审核

Brody I. · 已于 over 2 years前审核

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. · 已于 over 2 years前审核

LUCAS S. · 已于 over 2 years前审核

Sudharshan N. · 已于 over 2 years前审核

Alexey T. · 已于 over 2 years前审核

Suchita B. · 已于 over 2 years前审核

DANIEL M. · 已于 over 2 years前审核

Jayesh Arun B. · 已于 over 2 years前审核

ROBERTO O. · 已于 over 2 years前审核

Rahul M. · 已于 over 2 years前审核

David M. · 已于 over 2 years前审核

Harold H. · 已于 over 2 years前审核

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. · 已于 over 2 years前审核

Ming H. · 已于 over 2 years前审核

LUIS F. · 已于 over 2 years前审核

Kidus B. · 已于 over 2 years前审核