John Snow Labs Spark-NLP 3.1.1: New and improved ALBERT with support for external Transformers, real-time metrics in Python notebooks, bug fixes, and many more improvements!

[maziyarpanahi]

---

Overview

We are pleased to release Spark NLP 🚀 3.1.1! We have a new and much-improved ALBERT annotator with support for HuggingFace 🤗 models in Spark NLP. We managed to make AlbertEmbeddings almost 7x times faster on GPU compare to prior releases!

As always, we would like to thank our community for their feedback, questions, and feature requests.

---

New Features

• Migrate AlbertEmbeddings to TensorFlow v2. This allows the importing of HuggingFace ALBERT models to Spark NLP
• Migrate AlbertEmbeddings to BatchAnnotate to allow better performance on accelerated hardware such as GPU
• Enable stdout/stderr in real-time for child processes via sparknlp.start(). Thanks to PySpark 3.x, this is now possible with sparknlp.start(real_time_output=True) to have the outputs of Spark NLP (such as metrics during training) right in your Jupyter, Colab, and Kaggle notebooks.
• Complete examples for all annotators in Scaladoc APIs WIP: Doc Examples update #5668

---

Bug Fixes & Enhancements

• Fix YakeModel issue with empty token Fixing yake bug, was throwing exception on empty token input #5683 thanks to @shaddoxac
• Fix getAnchorDateMonth method in DateMatcher and MultiDateMatcher Bugfix/get anchor date month denormalized #5693
• Fix the broken PubTutor class in Python python pubtator fix #5702
• Fix relative dates in DateMatcher and MultiDateMatcher such as day after tomorrow or day before yesterday fix datematecher utils for relatives dates #5706
• Add isPaddedToken param to PubTutor python pubtator fix #5702
• Fix issue with logger inside session on some setup Fix logger issue in some cluster / add signatures to the sessions #5715
• Add signatures to TF session to handle inputs/outputs more dynamically in BertEmbeddings, DistilBertEmbeddings, RoBertaEmbeddings, and XlmRoBertaEmbeddings Fix logger issue in some cluster / add signatures to the sessions #5715
• Fix XlmRoBertaEmbeddings issue with init_all_tables Fix logger issue in some cluster / add signatures to the sessions #5715
• Add missing YakeModel from annotators
• Add missing random seed param to ClassifierDLApproach, MultiClassifierDLApproach, and SentimentDLApproach Missing getter/setter for random seed #5697
• Make the Java Exceptions appear before Py4J exceptions for ease of debugging in Python Make the Java error appear first #5709
• Make sure batchSize set in NerDLModel is the same internally to feed TensorFlow Use the same batchSize internally in NerDLModel #5716
• Fix a typo in documentation Fix typo in concepts.md #5664 thanks to @roger-yu-ds

---

Performance Improvements

Introducing a new batch annotation technique implemented in Spark NLP 3.1.1 for AlbertEmbeddings annotator to radically improve prediction/inferencing performance. From now on the batchSize for these annotators means the number of rows that can be fed into the models for prediction instead of sentences per row. You can control the throughput when you are on accelerated hardware such as GPU to fully utilize it.

Performance achievements by using Spark NLP 2.x/3.0.x vs. Spark NLP 3.1.1

(Performed on a Databricks cluster)

Spark NLP 2.x/3.1.0 vs. 3.0.x	CPU	GPU
ALBERT Base	22%	340%
Albert Large	20%	770%

We will update this benchmark table in future pre-releases.

---

Backward compatibility

We have migrated AlbertEmbeddings to TensorFlow v2, the earlier models prior to 3.1.1 won't work after this release. We have already updated the models and uploaded them on Models Hub. You can use pretrained() that takes care of it automatically or please make sure you download the new models manually.

---

Documentation

• HuggingFace to Spark NLP
• Models Hub with new models
• Spark NLP publications
• Spark NLP in Action
• Spark NLP documentation
• Spark NLP Workshop notebooks
• Spark NLP training certification notebooks for Google Colab and Databricks
• Spark NLP Display for visualization of different types of annotations
• Discussions Engage with other community members, share ideas, and show off how you use Spark NLP!

---

Installation

Python

#PyPI

pip install spark-nlp==3.1.1

Spark Packages

spark-nlp on Apache Spark 3.0.x and 3.1.x (Scala 2.12 only):

spark-shell --packages com.johnsnowlabs.nlp:spark-nlp_2.12:3.1.1

pyspark --packages com.johnsnowlabs.nlp:spark-nlp_2.12:3.1.1

GPU

spark-shell --packages com.johnsnowlabs.nlp:spark-nlp-gpu_2.12:3.1.1

pyspark --packages com.johnsnowlabs.nlp:spark-nlp-gpu_2.12:3.1.1

spark-nlp on Apache Spark 2.4.x (Scala 2.11 only):

spark-shell --packages com.johnsnowlabs.nlp:spark-nlp-spark24_2.11:3.1.1

pyspark --packages com.johnsnowlabs.nlp:spark-nlp-spark24_2.11:3.1.1

GPU

spark-shell --packages com.johnsnowlabs.nlp:spark-nlp-gpu-spark24_2.11:3.1.1

pyspark --packages com.johnsnowlabs.nlp:spark-nlp-gpu-spark24_2.11:3.1.1

spark-nlp on Apache Spark 2.3.x (Scala 2.11 only):

spark-shell --packages com.johnsnowlabs.nlp:spark-nlp-spark23_2.11:3.1.1

pyspark --packages com.johnsnowlabs.nlp:spark-nlp-spark23_2.11:3.1.1

GPU

spark-shell --packages com.johnsnowlabs.nlp:spark-nlp-spark23-gpu_2.11:3.1.1

pyspark --packages com.johnsnowlabs.nlp:spark-nlp-gpu-spark23_2.11:3.1.1

Maven

spark-nlp on Apache Spark 3.0.x and 3.1.x:

<dependency>
    <groupId>com.johnsnowlabs.nlp</groupId>
    <artifactId>spark-nlp_2.12</artifactId>
    <version>3.1.1</version>
</dependency>

spark-nlp-gpu:

<dependency>
    <groupId>com.johnsnowlabs.nlp</groupId>
    <artifactId>spark-nlp-gpu_2.12</artifactId>
    <version>3.1.1</version>
</dependency>

spark-nlp on Apache Spark 2.4.x:

<dependency>
    <groupId>com.johnsnowlabs.nlp</groupId>
    <artifactId>spark-nlp-spark24_2.11</artifactId>
    <version>3.1.1</version>
</dependency>

spark-nlp-gpu:

<dependency>
    <groupId>com.johnsnowlabs.nlp</groupId>
    <artifactId>spark-nlp-gpu-spark24_2.11</artifactId>
    <version>3.1.1</version>
</dependency>

spark-nlp on Apache Spark 2.3.x:

<dependency>
    <groupId>com.johnsnowlabs.nlp</groupId>
    <artifactId>spark-nlp-spark23_2.11</artifactId>
    <version>3.1.1</version>
</dependency>

spark-nlp-gpu:

<dependency>
    <groupId>com.johnsnowlabs.nlp</groupId>
    <artifactId>spark-nlp-gpu-spark23_2.11</artifactId>
    <version>3.1.1</version>
</dependency>

FAT JARs

• CPU on Apache Spark 3.x: https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/jars/spark-nlp-assembly-3.1.1.jar

• GPU on Apache Spark 3.x: https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/jars/spark-nlp-gpu-assembly-3.1.1.jar

• CPU on Apache Spark 2.4.x: https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/jars/spark-nlp-spark24-assembly-3.1.1.jar

• GPU on Apache Spark 2.4.x: https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/jars/spark-nlp-gpu-spark24-assembly-3.1.1.jar

• CPU on Apache Spark 2.3.x: https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/jars/spark-nlp-spark23-assembly-3.1.1.jar

• GPU on Apache Spark 2.3.x: https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/jars/spark-nlp-gpu-spark23-assembly-3.1.1.jar

---

This discussion was created from the release John Snow Labs Spark-NLP 3.1.1: New and improved ALBERT with support for external Transformers, real-time metrics in Python notebooks, bug fixes, and many more improvements!.