John Bauer comments

Results 1064 comments of


                                            John Bauer

Using newly trained model of clarkkev/deep-coref in Stanford neural coref system

As near as I can tell, there is a discrepancy in the features produced by the python and the java code. In python, the features becomes a 674 dimension vector,...

Using newly trained model of clarkkev/deep-coref in Stanford neural coref system

Also, random observation - did you notice it running on GPU, or did you see any flags etc which would put it on the GPU? My nvidia-smi shows my GPU...

Using newly trained model of clarkkev/deep-coref in Stanford neural coref system

To train the models with 668 dimensions instead of 674, there's a couple lines I changed in model_properties.py of deep-coref: - self.use_genre = use_genre + self.use_genre = False - self.single_size...

Using newly trained model of clarkkev/deep-coref in Stanford neural coref system

Just a small heads up, with the next version we're upgrading ejml from 0.38 to 0.39. This breaks the serialization of existing coref models... Internally we used a tool to...

Question/Feature :: NER tag overwriting if its identified as other from one classifier but however identified correctly with next classifier, Is it possible with stanford core nlp???

In general it will overwrite O, but not other labels. After all, there's no easy way to distinguish the earlier or later model

Question: Best practices for converting OntoNotes to UD

The PTB corrector was only intended for the PTB, not OntoNotes. You could always try diffing the two lines to see if there is any difference, and if so, if...

Question: Best practices for converting OntoNotes to UD

I have since updated the lemmatizer to incorporate many of the fixes in EWT, although it is still not 100% the same

Questions about Spanish model

I'm training one now for Stanza on the same data used in CoreNLP. I expect it to be more accurate with the Stanza model

Questions about Spanish model

Gotta process it tonight, then put it up tomorrow morning On Wed, Jun 8, 2022 at 12:35 PM Xuanci Zheng ***@***.***> wrote: > I'm training one now for Stanza on...

Questions about Spanish model

Oops! Looking through the old issues, I see that I never posted a followup, although the model has been there for quite a while now on the dev branch. If...