"Param poolingLayer does not exist" 在 spark-nlp 中加载 BERT 嵌入模型时出错

Question

我的 NLP 管道使用来自 johnsnowlabs 的预训练 BERT 嵌入模型“bert_base_uncased”。但是在加载这个下载的模型时，我遇到了以下异常。

Caused by: java.util.NoSuchElementException: Param poolingLayer does not exist.
    at org.apache.spark.ml.param.Params$$anonfun$getParam.apply(params.scala:729)
    at org.apache.spark.ml.param.Params$$anonfun$getParam.apply(params.scala:729)
    at scala.Option.getOrElse(Option.scala:121)
    at org.apache.spark.ml.param.Params$class.getParam(params.scala:728)
    at org.apache.spark.ml.PipelineStage.getParam(Pipeline.scala:42)
    at org.apache.spark.ml.util.DefaultParamsReader$Metadata$$anonfun$setParams.apply(ReadWrite.scala:591)
    at org.apache.spark.ml.util.DefaultParamsReader$Metadata$$anonfun$setParams.apply(ReadWrite.scala:589)
    at scala.collection.immutable.List.foreach(List.scala:392)
    at org.apache.spark.ml.util.DefaultParamsReader$Metadata.setParams(ReadWrite.scala:589)
    at org.apache.spark.ml.util.DefaultParamsReader$Metadata.getAndSetParams(ReadWrite.scala:577)
    at org.apache.spark.ml.util.DefaultParamsReader.load(ReadWrite.scala:497)
    at com.johnsnowlabs.nlp.FeaturesReader.load(ParamsAndFeaturesReadable.scala:12)
    at com.johnsnowlabs.nlp.FeaturesReader.load(ParamsAndFeaturesReadable.scala:8)
    at org.apache.spark.ml.util.MLReadable$class.load(ReadWrite.scala:380)
    at com.johnsnowlabs.nlp.embeddings.BertEmbeddings$.load(BertEmbeddings.scala:302)
    at com.johnsnowlabs.nlp.embeddings.BertEmbeddings.load(BertEmbeddings.scala)

Answer 1

很可能您有混合版本的模型和库。异常抱怨的参数最近已从 Bert 模型中删除。所以你应该尝试不同的预训练 Bert 模型，

bert_large_uncased or bert_base_cased

Answer 2

基于 spark-nlp slack 频道的帮助。我通过使用来自 spark-nlp 的最新训练模型解决了这个问题。对于 BERT，我使用了模型“bert_base_cased_en_2.6.0_2.4_1598340336670”

之前我使用的是2.4.0版本的模型，使用2.6.0版本的模型后，我没有看到任何错误。 poolingLayer 参数不再存在于新模型中。

"Param poolingLayer does not exist" 在 spark-nlp 中加载 BERT 嵌入模型时出错

"Param poolingLayer does not exist" error coming while loading BERT embedding model in spark-nlp

nlp

johnsnowlabs-spark-nlp