STREAM: Counting iterator exceeded timeout of 10 seconds after 69 tasks
17:54:10: CORS: initialized with wildcard "*" CORS origins
__ Starting the web server at http://0.0.0.0:8080 ...
Open the app in your browser and start annotating!
For information:
============================== __ Prodigy Stats ==============================
Version 1.11.7
....
Platform Linux-3.10.0-1160.25.1.el7.x86_64-x86_64-with-glibc2.31
Python Version 3.9.9
Database Name MySQL
Database Id mysql
Are you certain that dataset:foo contains data in the right format for the rel.manual recipe? Is it possible for us to have a single example for that dataset so that we may try to reproduce it locally?
Where ./data/schedule_participants.jsonl contains:
{"text":"Participantes:\n- Secretário-Executivo Adjunto do Gabinete de Segurança Institucional, Osmar Lootens Machado; e\n- Assessor Diplomático, Manuel Innocencio de Lacerda Santos Junior.\n"}
The only thing I can't share here is my model pt_agenda_ner_rel. However, if you email me I can send it to you.
It seems, on my end, that when I remove the --wrap flag it does work. It also seems that when I select "wrap" from the checkbox afterward that it breaks once again.
@vitorbaptista can you confirm you see the same behavior? I'm diving deeper into why this is happening but if you can confirm the same behavior on your end I can confirm that it's the same issue.
I've found something else too. It seems that your example contains a newline. If you remove it, so that the example looks like this:
{"text":"Participantes:\n- Secretário-Executivo Adjunto do Gabinete de Segurança Institucional, Osmar Lootens Machado; e\n- Assessor Diplomático, Manuel Innocencio de Lacerda Santos Junior."}
Then it all seems to work again. Could you confirm that removing the \n works?