Kubeflow

MLOps pipeline > List of MLOps tools

In this document we list a series of MLOps tools that are currently used in Charmed Kubeflow. All these tools are open source and free to use!

Contents:

• Kubeflow Katib
• Kubeflow Pipelines (KFP)
• Minio
• MLFLow
• Seldon Core

Kubeflow Katib

See also upstream: Kubeflow | Katib

Kubeflow Katib is a Kubernetes-native project for automated machine learning (AutoML). Katib supports hyperparameter tuning, early stopping and neural architecture search (NAS).

Katib is the project which is agnostic to machine learning (ML) frameworks. It can tune hyperparameters of applications written in any language of the users’ choice and natively supports many ML frameworks, such as TensorFlow, MXNet, PyTorch, XGBoost, and others.

Hyperparameters are the variables that control the model training process. They include:

• The learning rate.
• The number of layers in a neural network.
• The number of nodes in each layer.

Kubeflow Pipelines (KFP)

See also upstream: Kubeflow Pipelines (KFP)

Kubeflow Pipelines (KFP) is a workflow engine that allows us to specify tasks and their configuration, environment variables and secrets. Additionally, KFP ensures all tasks are correctly scheduled in the proper order of execution.

Minio

See also upstream: Minio

Minio is an object storage system that keeps the data safe and secured. Minio provides an AWS S3 compatible API and can work as a gateway for cloud storage or as a standalone object storage system. Minio allows you to have S3 buckets in your own data centre with high durability and availability.

MLFLow

See also upstream: MLFlow

MLFLow is an experiment and model repository that will help you track model training results, compare them and keep track of your deployed models. It tracks all the metadata about your models and experiments in a single place.

Seldon Core

See also upstream: Seldon Core

Seldon Core is a platform to deploy machine learning models on Kubernetes at scale as microservices. It supports REST and gRPC protocols, manual and auto-scaling. Thic can help to make deploying models a whole lot easier and faster too.