Desired output

[Miaoishappening]

Hi, I tried several models in the release. I am not so sure what should the desired output look like.

I used droc_incremental_no_segment_distance.mar as deployed model. I tried two input methods, that are showed in the README, the result confused me a little...

------> Input: model-server@2387c6518b8e:~$ http http://127.0.0.1:8080/predictions/droc_incremental_no_segment_distance output_format=conll tokenized_sentences:='[["Die", "Organisation", "gab", "bekannt", "sie", "habe", "Spenden", "veruntreut", "."], ["Next", "sentence", "goes", "here", "!"]]'

------> Output: HTTP/1.1 200 Cache-Control: no-cache; no-store, must-revalidate, private Expires: Thu, 01 Jan 1970 00:00:00 UTC Pragma: no-cache connection: keep-alive content-length: 950 x-request-id: 1f09b5b8-eb02-4fa3-b082-bc4046dc47a4

begin document

memory_file 1 1 Die - - - - - - - - (0 memory_file 1 2 Organisation - - - - - - - - 0) memory_file 1 3 gab - - - - - - - - - memory_file 1 4 bekannt - - - - - - - - - memory_file 1 5 sie - - - - - - - - - memory_file 1 6 habe - - - - - - - - - memory_file 1 7 Spenden - - - - - - - - - memory_file 1 8 veruntreut - - - - - - - - - memory_file 1 9 . - - - - - - - - -

memory_file 2 1 Next - - - - - - - - - memory_file 2 2 sentence - - - - - - - - - memory_file 2 3 goes - - - - - - - - - memory_file 2 4 here - - - - - - - - - memory_file 2 5 ! - - - - - - - - -

end document

------> Input: model-server@2387c6518b8e:~$ http http://127.0.0.1:8080/predictions/droc_incremental_no_segment_distance output_format=raw text="Die Organisation gab bekannt sie habe Spenden veruntreut."

------> Output: HTTP/1.1 200 Cache-Control: no-cache; no-store, must-revalidate, private Expires: Thu, 01 Jan 1970 00:00:00 UTC Pragma: no-cache connection: keep-alive content-length: 62 x-request-id: 11fc0cf2-4ba1-42cc-9d28-ca329050f6da

Die [Organisation gab]? bekannt sie habe Spenden veruntreut.

So my question is: Am i using the model correctly? Because the answer seems weird. Why there is a question mark in the text instead of a number?

[hatzel]

Sorry for the late answer. Could this perhaps be an encoding issue on your end? We use subscripts for this which depending on your terminal may cause issues.