DEV Community

Rishit Dagli
Rishit Dagli

Posted on • Originally published at twitter.com

đź’ˇTensorFlow Tip, use .cache

đź’ˇ#TensorFlowTip

Use .cache to save on some ops like opening file and reading data during each epoch

  • transformations before cache are only run for 1st epoch
  • can cache in-memory or on-disk
  • not repeatedly perform expensive CPU ops

See the speedup with .cache in this image.

Alt Text

Try it out for yourself:

Discussion (0)