IBM Launches Codeflare, An Open Source Machine Learning Workflow Catalyst

Analyzing simple, minute data or ephemeral event statistics maybe a walk in the park; datasets are another case. Data scientists and machine learning analytics have spread to farther depths in the world. As they grow larger, they become more complex, necessitating a significant amount of configuration work. Researchers spend less time doing data science and more time keeping their systems up to date, which can be challenging at times.

Up steps International Business Machines Corp., introducing a new solution that makes it easier to integrate, scale, and accelerate complicated multi-step analytics and machine learning pipelines on the hybrid multi-cloud.

CodeFlare is an open source machine learning structure that developers will relish that aims at unsophisticated workflows that can be built and managed shortly. Commonplace big data and AI workflow paradigm are that having 10,000 pipelines running would require 4 hours to yield results. CodeFlare drastically reduces that to a quarter of an hour.

Training and optimizing a machine learning model - a very demanding task, has now become simplified. These steps involving data sorting, feature extraction, and model optimization are imperative when designing a machine learning model. CodeFlare, with a Python-based interface, makes this process simpler and faster.

CloudFlare is built on Ray; a rising open source distributed cloud and machine learning framework. However, IBM claims CodeFlare has been modified to make scaling workflows easier, extending the rudiments of Ray.

CodeFlare pipelines run on IBM's new serverless platform IBM Cloud Code Engine and Red Hat OpenShift. This allows users to deploy CodeFlare almost anywhere, extending the benefits of serverless to data scientists and AI researchers, IBM said.

CodeFlare pipelines could run on Red Hat OpenShift and IBM's emerging serverless platform, IBM Cloud Code Engine, or any platform based on Kubernetes. These platforms have outlandish user statistics, which means CodeFlare could be deployed anywhere in the world. This also means data scientists and AU researchers can leverage serverless computing capabilities and integrate with other cloud native applications.

This also makes it easier to integrate and bridge with other cloud native ecosystems by giving adapters to event triggers such as the arrival of a new file and the ability to load and separate data from various sources, including cloud object repositories data lakes, and distributed file systems.

IBM foresees a future without data science complexities by providing a platform that evolves without limit. Data scientists will consistently delve deeper into competent tools, removing limiters on their research while CodeFlare and IBM's upcoming arsenal handle challenges that tackle configuration and deployment.