JohnSnowLabs/spark-nlp-workshop

Parameters that can be tuned for BertEmbeddings component

dsvrsecc opened this issue · 1 comments

Hi,

In this blog 'https://towardsdatascience.com/named-entity-recognition-ner-with-bert-in-spark-nlp-874df20d1d77' ,there is a parameter 'setPoolingLayer' ,but in the latest versions 3.1.3 ,I couldn't define this parameter as it is not available. In this case,what is the default value of pooling layer of bert that is set.

image

Hi @dsvrsecc, the latest version is not 3.1.3, it is 3.4.0 now. please try to download the latest version by using pip install sparknlp==3.4.0 and then try to use this parameter. If you still have an issue, please share it with us.