How to train a model on more than one label

Question

GabrielGhe opened this issue 4 years ago · 1 comments

Given the multi-class text analysis notebook, how would that work with a two-dimensional output layer?

input = { text: ['word1', 'word2'], category: 'article', sentiment: 'good' }

The input is the text and the labels are category and sentiment.
How would that work?

Answer 1 · 2020-06-22T14:15:42.000Z

I figured out that you can have/return multiple fc layers in the forward pass.