sinaraml 2023.12.25

SinaraML

SinaraML Framework is a framework for creating and operating MLOps platforms

From the first sight of a Data Scientist with Sinara to receiving the first Docker image with a model accessible via REST, it will take 15 minutes only.

SinaraML Framework provides Sinara Server, Sinara Storage, Sinara Lib, Sinara Step Template, Sinara Tools to the Data Scientist.

• SinaraML Server is a Jupyter Server, with all the necessary libraries for working with data and training models. SinaraML provides three Basic Servers for different purposes — classic ML, computer vision (CV) and natural language processing (NLP).
• SinaraML Storage is a long-term storage where ML pipeline stored input and output entities. Depending on infrastructure Sinara Storage can be implemented based on S3, HDFS protocols or local disk.
• SinaraML Spark (pandas_api) is a effective and uniform way to work with big data sets using Apache Spark. Pandas_api gives Apache Spark's effectivenes and familiar pandas datasets functions.
• SinaraML Lib is a compact library that contains everything you need to create ML pipelines, for data preparation and versioning, model versioning and serving.
• SinaraML Step Template is a component template for creating a SinaraML Step — an ML pipeline step. The ML pipeline consists of several steps. Each step based on this template.
• SinaraML CLI is a number of CLI tools for creating, deleting, stopping and starting a Sinara Server, creating docker images from BentoServices created by ML pipelines, ML pipelines management, visualization etc.
• SinaraML Basic is a preconfigured personal MLOps platform working on desktop, remote virtual machine which can be located on-prem or on clouds like Google Collab/DataProc + Google Objecs, Azure DataBricks + Azure Blobs, Yandex DataSphere/DataProc + S3 etc.
• SinaraML Customizable Infra is a way to customize orchestration of Sinara Server, Sinara Storage and Sinara Spark for integration with your infrastructure (Git, Docker repos, authentication and authorization methods including Active Directory).
• SinaraML Customizable Dev Process is a way to configure Sinara Template and Sinara Step for your development process.
• SinaraML Examples is a library of ready to use configurable ML pipelines can be customized for your needs.
• SinaraML Book allows you to dive deeply into the development of ML products with SinaraML examples.

To start you off, go to Getting started page to try SinaraML Tutorials

Installation

To install SinaraML CLI into your environment, run:

pip install sinaraml

Reload shell or reboot your machine after installation to enable CLI commands

Quick Start

Commands start with the keyword sinara (similar to git, docker, kubectl)
If a command call is made without a mandatory parameter, help is displayed on the available parameters and methods of calling the command, for example:

sinara server create

sinara server start

Or, for a remote VM platform:

sinara server create --platform remote_vm

sinara server start

To remove a server, run:

sinara server remove