DL_ConfigureLogging


Header: AVLDL.h
Namespace: avl
Module: DeepLearning

Filter is used to set minimum level of logged messages.

Syntax

void avl::DL_ConfigureLogging
(
	avl::DeepLearningLogLevel::Type inMinLogLevel
)

Parameters

Name Type Default Description
Input value
inMinLogLevel DeepLearningLogLevel::Type Info Sets minimum level of logged messages

Description

This tool may be used to control amount of logs generated by other Deep Learning tools. By default, only Error, Warning and Info messages are logged. In a typical operation it may generate several megabytes of data. One file per day is created. Only last 3 files are stored - the rest is deleted to contain size of log files.

If such disk space consumption is unacceptable, inMinLogLevel may be set to Warning or even Error. This will ensure that subsequent calls to Deep Learning functions will write substantially less data (if anything). However, it will not lead to any observable performance improvement.

Using Debug or Trace leads to much larger log files and worse performance. It should be used only to diagnose problems.

Log files are saved to a directory pointed by a AVLDL_LOGS_DIR environment variable. If it is not set at the moment of loading AVLDL.dll or it is incorrect, no log files will be written. If this path contains a nonexistent directory, Deep Learning Library will try to create it.

This tool is safe to call from multiple workers or threads. However, it is advisable to call this tool at the beginning of each worker/thread using Deep Learning tools.

Please note, that this mechanism is completely independent of the logging system in Adaptive Vision Studio or Adaptive Vision Library.