How to use TF_XLA_FLAGS on Xavier in NVIDIA Machine Learning for Jetson/L4T Container?

How to use XLA using such flags on Xavier in machine learning container?

Instructions specify:
TF_XLA_FLAGS=--tf_xla_auto_jit=2 path/to/your/tf/program

what is the path/to/your/tf/program ?
The path to the TF installation? or the python program that contains tf. commands?

An example would be good. Thanks.


This is an execution command.

The path/to/your/tf/program is the script of your tensorflow application.
For example.

$ TF_XLA_FLAGS=--tf_xla_auto_jit=2 /home/nvidia/