Michael Shtelma

Command execution result does not always have the "data" field.

2

## Proposed changes Command execution results do not always contain the "data" field. Added null check. ## Types of changes What types of changes does your code introduce to dbx?...

Pre-trained model for NER

1

Hi, is there a pre-trained model for the NER task?

Different update statements on column tables lead to AssertionError: java.lang.AssertionError: "assertion failed: Unexpected nulls=2 for withHeader=false”

2

The following statement in spark-shell "spark.sql(“update position set VAL115=VAL115*2, version=4 “).collect" leads to the mentioned in the header error. The same assertion error occurs after many other update statements

Hi, is there somewhere a model pre-trained for BertForTokenClassification? I can of course fine-tune the existing model for NER, but how can I download the whole dataset including pubmed? Thanks!