Text Classification Custom Label Issue

,
1

I'm having a bit of trouble with the text classification feature of Prodigy.
Recently, I created my own .jsonl dataset of news article texts to be manually classified into custom labels/categories. There are 20 lines of texts with their sources. For example:

{"text":"One of the reasons Trump lost his re-election bid is that the American people were tired — not only of the daily chaos and his non-stop tweets, but they knew he made up a lot of things and they lost trust in him.","meta":{"source":"The Hill"}},

After creating this dataset, I put the following code in the terminal:

prodigy textcat.manual news_articles ./article_texts.jsonl --label Honesty,Economic Management,Relations,Integrity

However, it failed to open and it didn't work.
I feel like I might be doing something wrong and I would love if to get some insight on how to fix my problem. :face_with_monocle:

2

Have you tried ‘quotes’ around the label with a space in it?

3

Can you share some more details on what exactly doesn't work? Do you see an error message?

4

Well, when I run the code, it opens up the browser, but all I get is a blank page with this text:

#!/bin/sh
python -m prodigy "$@"

What do you think?

5

@Dawudis Are you on Windows? It looks like your shell can't interpret the shortcut prodigy. If you use python -m prodigy instead, it should work :slight_smile:

6

Yup, that did the trick! Really appreciate the help!

1 Like