PMLE Question 81

Community Discussion

Written by other learners. Comments can disagree with the marked answer — check the official documentation when in doubt.

• Chose C1

  C is easiest and least effort so should be tried first.

• Chose C1

  I focus on 'the first step' - splitting CSV and interleave are very easy and quick to try/

• Chose C1

  By splitting the data into multiple CSV files and using tf.data.Dataset.interleave(..., num_parallel_calls=tf.data.AUTOTUNE) you: -Enable parallel reads from Cloud Storage -Dramatically improve throughput -Reduce GPU/CPU idle time -Achieve performance gains with minimal engineering effort