AI Units Consumption

Help AI Center
,
1

I have created a ML Skill to extract the data from the invoices and selected the below option

image
image345Ă—276 8.72 KB

I have also developed a code to extract the data from invoices by using Document understanding packages.

My query related to this is : if we selected this option to undeployed the skill after 1 Day of inactivity, then if I run the code to extract data, will it work without manually enabling the ML Skill in UiPath cloud?

Can anyone help me with the detail working this?

2

Hi @rnahasnahasuddin

As per my knowledge,
if you have selected the option to undeploy the ML Skill after 1 day of inactivity, you will need to manually enable it in the UiPath Cloud if you want to run your code for extracting data from invoices after that period. It’s essential to manage the deployment settings to ensure seamless operation of your automation processes.

If you found helpful, mark as a solution
Happy Automation!

3

Hey @rnahasnahasuddin
If you select the option Undeploy Skill after 1 Day of inactivity, your ML Skill will automatically be disabled if it’s not used for 24 hours. When that happens, any process which use the skill won’t work and will throw an error because the skill is not active.
To make it work again, you’ll need to manually enable the skill in AI Center.

The image shows instructions for redeploying stopped ML Skills in the Automation Suite by clicking the "Resume" button on the ML Skills page. (Captioned by AI)
The image shows instructions for redeploying stopped ML Skills in the Automation Suite by clicking the "Resume" button on the ML Skills page. (Captioned by AI)957Ă—722 83.2 KB

4

So if we select the below option as Never

image
image347Ă—236 7.51 KB

then AI units will be consumed hourly, if it is inactive also, right?

5

Hi @rnahasnahasuddin

With the “Never” setting, your ML Skill will be ready for extraction tasks at any time, eliminating the need for manual intervention.

6

@rnahasnahasuddin
If you select the “Never” option for undeploying the ML Skill, the skill will remain deployed at all times, even when it’s not actively used. While this means the skill will be instantly available when needed, it also result in continuous hardware consumption

The image outlines the hardware costs for deploying ML skills and pipelines, including the unit costs for various hardware specifications and the impact of replicas on availability and cost computation. (Captioned by AI)
The image outlines the hardware costs for deploying ML skills and pipelines, including the unit costs for various hardware specifications and the impact of replicas on availability and cost computation. (Captioned by AI)1016Ă—948 57.9 KB

That could lead to unnecessary AI Units or hardware consumption over time. For optimized usage and cost control, it’s recommended to use auto-undeploy after inactivity when possible.

1 Like
8

Hey,
As the Ai Unit consumption has been a big problem for us in the organization, i’ve created an API connection through an external application to deploy/stop the Skill when the process starts running. that might help you control the consumption.

1 Like
9

@rnahasnahasuddin,

Having a fixed schedule to enable and disable the ML skill is the best possible solution here.

10

How can we schedule to enable the ML Skill?

11

@rnahasnahasuddin,

No native approach provided. You will have to build it using API like @daniela.rosenstein mentioned.

12

Hi @rnahasnahasuddin

Please check below official link,