This is a very simple standalone console app exercising the oneDAL extensions to ML.NET.
The oneDAL library takes advantage of features on modern CPU processors, specifically Intel-x64 variants, so you'll need one of those.
To run this sample, you'll also need .NET 8.0..
With this, trying this out should be as simple as
dotnet runIn some instances, you may see an error loading a library. In that
case, please take a look at the script run.sh, that sets the
LD_LIBRARY_PATH environment variable on Linux (on Windows you may
want to do the same with PATH)
The "bicep" file included provisions an IceLake machine on Azure running Linux, which you can access using the credentials you specify:
RGNAME=<some resource group name>
az group create --name $RGNAME --location 'westus2'
To use a fresh ssh key, the following method works:
ssh-keygen <ssh options> -C "<username>@azure" -f <keypair>
az deployment group create --resource-group $RGNAME --template-file infra.bicep --parameters vmname=<vmname> vmuser=<username> publickey="$(< <keypair>.pub)"
az deployment group show -g $RGNAME -n infra --query properties.outputs.sshCommand
Note that you may want to use a more useful name than the name of the template for the deployment, for easier management.
Log on to the VM using the credentials you provided at instantiation:
ssh -i <keypair> <username>@<output_of_show>
Clone this repo and then issue
git clone --depth=1 https://github.com/rgesteve/tryprev3.git
run.sh
after cloning this repo on the VM.
Please let us know of any problems you experience when trying to run this, or any suggestions for improvements at the ML.NET repo. We look forward to your feedback!