Turkish-#NLP: Anyone interested in a Turkish BERT and wants to evaluate it on downstream tasks?
I did evaluation only for UD PoS tagging - any help is really appreciated!
Would really like to have a proper evaluation before adding it to the @huggingface Transformers hub
-
-
Well upload speed depends on your connection I guess, but depending on the model size, things might take hours... I’d go for local repo for probing -https://github.com/UKPLab/linspector-web …
-
It’d be great if you could send me a snippet/log and maybe the other problems you’ve encountered - there are exciting optimizations (and possible exciting integrations) coming to scale things up and I’d like to be sure to address all user problems (sahin@ukp.tu-darmstadt.de)
Kraj razgovora
Novi razgovor -
-
-
You can choose “contextual” -“Turkish”- and the probing tasks (case marking, pos, tense) you’d want to test. Then choose the type of Transformer model and upload it. It’ll tell you how it performs. It may be really slow now, but it’s open source, you can deploy it your own
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
I also have a “considerably” large Turkish text compilation— however may not be very balanced (well i bet you can’t find a large balanced corpora anyway) Just send me an email
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
This looks pretty awesome
I just uploaded an archive (with pytorch_model.bin, vocab.txt and config.json) but the interface still shows "uploading" - I'll wait a bit now. Should I try the GitHub repo
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.