Add support for petastorm

Description

Petastorm is a library designed for doing deep learning on large datasets in HDFS stored in Parquet format, this library would be useful for Hopsworks users. It can also be integrated in the feature store.

Petastorm depends on several libraries, such as opencv, libhdfs3, and tensorflow, which needs to be configured in Chef to avoid dependency conflicts with existing libraries.

Assignee

Kim Hammar

Reporter

Kim Hammar

Labels

None

Fix versions

Priority

Medium
Configure