LunchBox ML Neural Network Inputs

Hey everyone,

I am trying to write a definition that uses Lunchbox ML to name rooms.
For training and testing data I was going to use area and perimeter.
I rebuild the Grasshopper definition that Proving Ground provides on their website.
Dynamo is very explict about the type of input for the Neural Network and unfortunately I do not understand enough programming language to know what the error message is telling me.

Error Message in Networks.NeuralNetwork:
Warnung:Dictionary.ValueAtKey erwartet Argumenttyp(en) (Dictionary, string), wurde jedoch mit (Function, string) abgerufen.

Hope you understand the german error code :slight_smile:

What do I do wrong with the inputs?

Here’s the script:

Thanks in advance!

Oh, btw the training data section is a bit messy, some unused nodes from trying different things. Sorry

shameless self bump

Hi, don’t you have to create a Lunchbox Dataset from the Training data as well? And not just go directly from excel? It would really help to get script from you so that we don’t have to reproduce everything. :slight_smile:


Thanks, I’ll upload the script later!
Can you explain the difference between Dataset and Datatable to me?


a dataset is a collection of datatables.

Here you go, thanks for your effort!

Raumliste_Hochhaus.xlsx (28.6 KB)
RoomNaming_MachineLearning.dyn (54.5 KB)

anyone else playing with LunchboxML?