Tensorflow Data API: prefetch when training on CPU?

Asked Aug 11 '18 at 06:14

Active Aug 11 '18 at 06:14

Viewed 65 times

I'm training on a MacBook pro (no GPU). After watching https://www.youtube.com/watch?v=SxOsJPaxHME I saw that I could improve my input pipeline by reordering some ops and adding num_parallel_calls.

Question:

Since I'm not using a GPU, does it make sense to include dataset.prefetch(1) to the end of my pipeline?

edited Jun 20 '20 at 09:12

Community

asked Aug 11 '18 at 06:14

rodrigo-silveira

12,607
11
69
123

Probably not. Have you measured any difference on your Mac book? – P-Gn Aug 11 '18 at 06:30
Have not benchmarked hoping someone had already done that work. Guess I know what I'll be doing today... – rodrigo-silveira Aug 11 '18 at 14:57

Tensorflow Data API: prefetch when training on CPU?

Question:

0 Answers0