John Snow Labs Spark-NLP 4.2.2: Support DBFS, HDFS, and S3 for importing external models, unifying LightPipeline APIs across supported languages for Image Classification, new fullAnnotateImage for Scala, new fullAnnotateImageJava for Java, support LightPipeline for QuestionAnswering pre-trained pipelines, and bug fixes #12996
maziyarpanahi
announced in
Announcement
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
📢 Overview
Spark NLP 4.2.2 🚀 comes with support for DBFS, HDFS, and S3 in addition to local file systems when you are importing external models from TF Hub and Hugging Face, unifying LightPipeline APIs across Scala, Java, and Python languages for Image Classification, the new fullAnnotateImage for Scala, the new fullAnnotateImageJava for Java, the support for LightPipeline for QuestionAnswering pre-trained pipelines, and bug fixes.
Do not forget to visit Models Hub with over 11400+ free and open-source models & pipelines. As always, we would like to thank our community for their feedback, questions, and feature requests. 🎉
⭐ New Features & improvements
fullAnnotate
inLightPipeline
for the path of images in ScalafullAnnotate
method inPretrainedPipeline
for ScalafullAnnotateJava
method inPretrainedPipeline
for JavafullAnnotateImage
toPretrainedPipeline
for ScalafullAnnotateImageJava
toPretrainedPipeline
for JavafullAnnotate
method inPretrainedPipeline
Predicted Entities
to all Vision Transformers (ViT) models and pipelinesBug Fixes
annotatorType
name in Python and Scala for Spark schema in Annotation, AnnotationImage, and AnnotationAudioRecursiveTokenizer
annotator affecting downstream NLP tasks in the pipeline📓 New Notebooks
📖 Documentation
Installation
Python
#PyPI pip install spark-nlp==4.2.2
Spark Packages
spark-nlp on Apache Spark 3.0.x, 3.1.x, 3.2.x, and 3.3.x (Scala 2.12):
GPU
M1
Maven
spark-nlp on Apache Spark 3.0.x, 3.1.x, 3.2.x, and 3.3.x:
spark-nlp-gpu:
spark-nlp-m1:
FAT JARs
CPU on Apache Spark 3.x/3.1.x/3.2.x/3.3.x: https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/jars/spark-nlp-assembly-4.2.2.jar
GPU on Apache Spark 3.0.x/3.1.x/3.2.x/3.3.x: https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/jars/spark-nlp-gpu-assembly-4.2.2.jar
M1 on Apache Spark 3.0.x/3.1.x/3.2.x/3.3.x: https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/jars/spark-nlp-m1-assembly-4.2.2.jar
AArch64 on Apache Spark 3.0.x/3.1.x/3.2.x/3.3.x: https://s3.amazonaws.com/auxdata.johnsnowlabs.com/public/jars/spark-nlp-m1-assembly-4.2.2.jar
What's Changed
Contributors
@galiph @agsfer @pabla @josejuanmartinez @Cabir40 @maziyarpanahi @Meryem1425 @danilojsl @jsl-builder @jsl-models @ahmedlone127 @DevinTDHa @jdobes-cz @Damla-Gurbaz @Mary-Sci
New Contributors
Full Changelog: 4.2.1...4.2.2
This discussion was created from the release John Snow Labs Spark-NLP 4.2.2: Support DBFS, HDFS, and S3 for importing external models, unifying LightPipeline APIs across supported languages for Image Classification, new fullAnnotateImage for Scala, new fullAnnotateImageJava for Java, support LightPipeline for QuestionAnswering pre-trained pipelines, and bug fixes.
Beta Was this translation helpful? Give feedback.
All reactions