Prevent TF from logging info messages during data collection (google#93)

Currently, whenever a TF model gets loaded in by LLVM in the
development mode advisors, TF outputs a slew of info messages
to STDOUT that aren't particularly useful. This patch gets
rid of those messages by setting an environment variable when
running the compilation commands that only enables warnings
and errors to be printed to the console. This makes it much
easier to see progress/results in several of the scripts within
this repository such as train_locally.py and
generate_default_trace.py when grabbing performance data for a
policy.

Author: boomanaiden154

compiler_opt/rl/compilation_runner.py

@@ -17,6 +17,7 @@
 import abc
 import dataclasses
 import json
+import os
 import subprocess
 import threading
 from typing import Dict, List, Optional, Tuple
@@ -30,6 +31,8 @@
 _COMPILATION_TIMEOUT = flags.DEFINE_integer(
     'compilation_timeout', 60,
     'Max duration (in seconds) after which we cancel any compilation job.')
+_QUIET = flags.DEFINE_bool(
+    'quiet', True, 'Whether or not to compile quietly (hiding info logging)')
 
 
 def _calculate_reward(policy: float, baseline: float) -> float:
@@ -152,8 +155,14 @@ def start_cancellable_process(
     TimeoutExpired: if the process times out.
     ProcessKilledError: if the process was killed via the cancellation token.
   """
+  command_env = os.environ.copy()
+  # Disable tensorflow info messages during data collection
+  if _QUIET.value:
+    command_env['TF_CPP_MIN_LOG_LEVEL'] = '1'
   with subprocess.Popen(
-      cmdline, stdout=(subprocess.PIPE if want_output else None)) as p:
+      cmdline,
+      env=command_env,
+      stdout=(subprocess.PIPE if want_output else None)) as p:
     if cancellation_manager:
       cancellation_manager.register_process(p)