Dear Community, I was very inspired in Ignite ML and I wanted to try it with Python. Particularly I was interested in compares Ignite ML VS Spark ML However, I came across the fact that pyignite component allows only to perform basic cache operations through the API and it has nothing to do with Ignite ML.
I have discussed this issue with Alexey Zinoviev <zaleslaw....@gmail.com> and he suggested to describe here all required features which are not presented now in Ignite. Therefore the list of required features: * Ignite ML and pyignite integration. Ignite ML is a fairly versatile ML, just inside driving on Ignite primitives, so Ignite ML and pyignite compatibility requires a lot of java code using py4j library to wrap Ignite ML with python. Also, I'm sure lots of python developers will be appreciated opportunity to test this solution in their tasks. * Ignite ML and PySpark integration. The really interesting case is using pyignite ML with data preprocessed via pyspark. As soon as I know the current version of Ignite supports only integration with Spark (not Pyspark) I hope I wrote this letter in accordance with the rules of the community. Also, I hope these cases will be interested in the dev community. BG Andrei Gavrilov. Software Engineer. EPAM Systems -- Sent from: http://apache-ignite-developers.2346864.n4.nabble.com/
