The CS_AUTOTUNE#

The Cerebras AUTOTUNE (CS_AUTOTUNE) is similar to the TensorFlow tf.data.AUTOTUNE. When you are targeting the CS system, using CS_AUTOTUNE instead of tf.data.AUTOTUNE will result in a better specification of parameters such as:

  • num_parallel_calls

  • cycle_length

  • num_parallel_reads

Use CS_AUTOTUNE

We recommend that you use CS_AUTOTUNE in place of tf.data.AUTOTUNE.

Using CS_AUTOTUNE#

Before using CS_AUTOTUNE, make sure to import it in your code as follows:

from cerebras.tf.tools.analyze_input_fn import CS_AUTOTUNE