hi @Clausia!
Thanks for your question and welcome to the Prodigy community
It looks like you simply added commas (,
) after each argument that aren't needed. You also added a colon :
at the end too.
You should try:
!prodigy rel.manual ner_news_headlines blank:en ./news_headlines.jsonl --label PERSON
Make sure you have news_headlines.jsonl
in whatever local folder you're running that command. If you don't know, you can run ls
command to see what files are in your local folder.
That's great. You may have already seen them, but we have docs on input format types.
The standard data input format for .txt
files is simply each document separated by a new line:
This is a sentence.
This is another sentence.
Alternatively, for .csv
, you can have:
Text,Label,Meta
This is a sentence.,POSITIVE,0
This is another sentence.,NEGATIVE,0.1
But to save you a lot of headaches in the future, I would strongly recommend trying to learn a little of .jsonl
manipulation. .jsonl
is a very common file type not just for Prodigy but for a lot of data, especially large data.
One of the easiest ways is to use pandas
like this:
Hope this helps!