Create Your First Machine Learning Pipeline in ML.NET - 2020 Update

[@Deetoya]

Thank you so much for the video. I would like to see more of this level stuff... This is really helpful for people who want to get into ML and are form .Net background. Thank you so much again for this.

[@JonWood]

Thanks! Glad you enjoy the videos! :)

[@SaschaDittmann]

Thanks for the updated version of your one year old video! Looking forward to you next one.

[@nazartvm]

Thanks... Can you explain that pipeline fully?

[@josbexerr5166]

Muchas gracias Mister Jom....muy didactico

[@tekimotod6277]

Can you show how to load a pipeline while loading from a database and using FieldAwareFactorizationMachine?

[@jan7477]

Hello, thank you for your both tutorials. Can i somehow save the knowledge of my AI? Because i'm doing a game and want to train the NPC-Enemy only one time not at every start-up

[@JonWood]

Sure! You can save the model to a file by using the context.Model.Save method.

[@13panagiwtis13]

Thanks for this helpfull video. Is it possible to get the R-squared (or anything that gives the quality) of the specific prediction ? I mean not from the evaluation of the whole model, but from every prediction, to indicate if it may be accurate or not.

[@VKjkd]

Exactly what I’ve been trying to do. For key drivers analysis you want the values at the predictor level. I see there is a calculate PFI function..

[@Vu_Van_Tiep]

thank you.how many do your visual studio intall version ?

[@JonWood]

You mean what version of visual studio I’m using?

[@vuhoangdung]

can you make an update for F# also?

[@JonWood]

There is this video for F# - kzbin.info/www/bejne/hoetfotjnqljpNE. But I'll go back through it and if there are any changes I'll create an update video for it :)

[@spacefishaviation276]

System.ArgumentOutOfRangeException: 'Could not find input column 'Response' (Parameter 'inputSchema')' pls help

[@JonWood]

Are you using the same data from the video?

[@spacefishaviation276]

@@JonWood No but I figured it out..

[@tesla_tower]

Thank you.

[@asgharnafemerkid8590]

Thanks alot

[@CristianE-bb3jm]

Where do I add layers and nodes?

[@JonWood]

That’s for creating neural networks from scratch which ML.NET doesn’t quite have the ability to do

[@jorge1869]

I love C #, but it is still very verbose compared to Python for ML. Microsoft has to continue improving this feature, it is not yet convincing. Regards

[@JonWood]

True, it doesn't have as many features as what you would get in frameworks in Python. Is there anything specific you need from it to be useful? Definitely feel free to add an issue for any feature requests - github.com/dotnet/machinelearning - or I'll be glad to relay them to the team, as well :)