AI Fabric On-premise GA

[Jeremy_Tederry]

Today, we are excited to announce the launch of the long-awaited AI Fabric – ON PREMISE in general availability! This will be our first fully-featured version of AI Fabric on-premise which is a single node configuration. With this launch, customers can deploy, run and improve machine learning models locally on VM instances provided by cloud providers such as AWS, Azure, and GCP. Installation on bare metal machines is not yet officially supported. With this release, customers enjoy the deployment flexibility of on-premise or the cloud (launched last June).
Along with the UiPath AI starter models and our new Academy course on AI Fabric, getting started with AI has never been easier. Get your customers to start their AI use cases quickly without requiring advanced data science skills!

AI Fabric on-premises also part of Enterprise trial

AI Fabric on-premises is available through the UiPath Enterprise on-premises trial for 60 days. You will need to fill out the trial form as is already the case. When you receive your trial email soon after, it will have a new section that contains a license file download URL and a link to documentation on how to install AI Fabric. Installation documentation can be found here.

New Content
· AI Fabric Academy Training Course
· AI Fabric Whitepaper
· AI Fabric Webinar
· AI Fabric Webpage (added on-premise trial, webinar, and academy course)

What’s next?

• Deploy Multi-node AI Fabric to Azure AKS
• Deploy Multi-node AI Fabric to Docker Enterprise
• DU Data Manager on AI Fabric

[mukeshkala]

Wow , This would be a very good add on , Most of the customers were not comfortable with sending Data outside organisation. With On- Premise AI Fabric , This will be awesome.

[Mr_Meeseeks]

Holy Guacamole this is big (as usual)

[Maneesha_de_silva]

@Jeremy_Tederry
Awesome
This is going to be huge,
Waiting for the testing

[Kesavaraj_K]

@Jeremy_Tederry
This is great news!

In Community Free Trail, Does AI Fabric have an GPU license if i wanted to train and host the python pickle schema model?

[Jeremy_Tederry]

There is no community free trial, only enterprise free trial for now. It is not including GPU for SaaS but it is for on prem (I mean you can provide a GPU on the machine and you are free to use it) but you can retrain of CPU also.

[Mr_Meeseeks]

Random Q- Whats your favorite GPU?

asking for a friend

[rlatta1]

Are there plans for this to be able to run on an on premises Computer Vision server?

[ashok_s]

Hi @Jeremy_Tederry,

We are planning to try AI Fabric enterprise trial and we would like to install on-prem AiFabric on our testing laptop which has Ubuntu 18.04 LTS on it. It has GPU and other hardware requirements mentioned on the installation guide. I just want to confirm with you if it is possible to install in the laptop or it has to be an VM provided by popular provider?

Thanks,
Ashok

[Sandeep_P]

@Jeremy_Tederry! Thanks for the announcement. So excited to start the journey with AI fabric

[neeraj.gupta]

hey this sounds great… But how is it on-prem if …“Installation on bare metal machines is not yet officially supported.”

[Calvink]

Echoing few comments here, I’m also wondering why is it called ON-PREMISE when you can only run it on servers/VMs purchased from AWS, Azure etc. We are still installing AIF on a cloud server, only difference is it’s self-managed compared to the first cloud native version released in June which is UiPath-managed. Am i accurate on this or is there another explanation?

I’m no expert in AWS/cloud servers but is it even possible to have local instances of these cloud servers completely in the client’s own premise? As there are still clients out there which would want 100% on-premise without touching any cloud components. Hope someone can clarify…

[Jeremy_Tederry]

Hi all
Sorry for delayed answer, we are now installing AIF outside of cloud VM. It’s still not what we recommend (as our process is more tested on those VM) but this is possible. Also as part of our last patch release couple of weeks ago we are supporting other linux distributions (RHEL and CentOS).
The main difference we have between what we call Cloud (SaaS) and on prem is who is responsible for the architecture. On SaaS we take care of everything, on prem it is your responsibility, whether it’s on your own servers or on a Cloud provider we call it on prem.
It also has an impact on which Orchestrator you can use. We are not supporting hybrid installation between SaaS and on prem so both (Orchestrator and AI Fabric) need to be on same base.

Let me know if that helps clarifying things.
Jeremy