combust/mleap

Integrate with spark-nlp

Open

#292 创建于 2017年10月21日

在 GitHub 查看
 (1 评论) (0 反应) (0 负责人)Scala (1,461 star) (312 fork)batch import
enhancementepichelp wanted

描述

spark-nlp

Is a library that was developed by john snow labs to offer common transformers for NLP to Spark. It would be incredibly useful to be able to train these models and then deploy them with MLeap.

https://github.com/JohnSnowLabs/spark-nlp

We still need to outline the work in terms of which transformers and models we need to offer support for and if we will include them in core modules or create separate modules for NLP.

贡献者指南

技术栈
scalaspark
领域
datamachine learning
议题类型
feature
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
5
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
over 1 week
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
stale
清晰度议题是否清楚说明期望改动、验收标准和下一步。
needs investigation
前置要求
ScalaApache SparkNLP basicsMLeap internals
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
15
研究方向
Investigate the current MLeap module structure in the combust/mleap repository, particularly how existing transformers are defined and serialized. Review the spark nlp library by John Snow Labs (github.com/JohnSnowLabs/spark nlp) to identify which transformers and models are most commonly used. Then propose a plan to either extend core MLeap modules or create separate NLP specific modules, referencing specific files such as MleapTransformer.scala or similar. The discussion in the issue indicates that the scope needs to be defined first.