====Cognition Training Times File====

The training time file for a Cognition AI is a simple text file that lists the different training times the system should learn from.

The file is called **trainingtimes.csv** will be located inside the folder you nominate in the //path// of your [[cognition_file_format|Cognition JSON file]].

Each line of the file contains a //time// or //times//, an optional set out //outputs// and an optionally an //asset//.

===Times===

The training time can be given as a single date and time in **YYYY-MM-DD HH:mm:ss** format, or if using a range it can be written as two times, separated by ' to '.

<wrap em>2022-11-18 02:52:26</wrap> and \\
<wrap em>2022-11-18 02:52:26 to 2022-11-18 02:52:42</wrap> \\

are both valid time ranges.

===Output Data===

In some cases, the outputs don't come from ARDI data. In this case, you need to give the output data values that are attached to the different training times.

To do that, add a hash (#) at the end of the line, followed by a comma-delimited list of output values.

For example, if you had two outputs (//OK// and //Failed// in that order), and the machine was working fine at lunchtime at the first of October, your line would look like this...

<wrap em>2022-11-18 12:00:00#1,0</wrap>

===Retargeting===

If you're trying to make a generic AI, you'll sometimes want to combine events across multiple assets. You can use training data from other assets by assigning an asset ID to the training time line.

Add an exclamation mark at the end of the line, followed by the asset ID of the ARDI asset the training values should come from.

For example, if you had training data coming from asset #147, your line should look like this...

<wrap em>2022-11-18 12:00:00!147</wrap>