By Catalog from Coalesce (https://coalesce.io)
The considerable volumes of data handled by modern organization and complex pipeline orchestration processes make productionizing ML models difficult and costly.
Feature stores are tools that emerged as a solution to these issues. This new trend is not going to stop, and we'd rather bring visibility and structure soon.
At CastorDoc, we believe the first step to structure the feature stores tools market, is more transparency. For that reason, we put up a list of all the feature stores tools we heard of.
Get started with Feature Store
Name | Link | Demo | Deployment | Cloud supported | Feature definitions | Automated transforms | Sharing and discovery | Monitoring | Security | More info |
|---|---|---|---|---|---|---|---|---|---|---|
Tecton | www.tecton.ai | https://www.youtube.com/watch?v=u_L_V2HQ_nQ | Fully managed cloud service | AWS | Available | Available | Feature versioningdependency management | mlops.community | ||
Feast | feast.dev | https://www.youtube.com/watch?v=vMreZGyYrh8&feature=emb_imp_woyt | Open-source | AWSGCPAzureOn-prem | Not available | Not available | Feature versioning | Not available | Data remains in end-user's cloud accountSSOData encryption at rest | mlops.community |
Hopsworks | www.hopsworks.ai | https://youtu.be/kadNWkT4gGM | Open-sourceFully managed cloud service | AWSAzureGCPOn-prem | Available | Available | Feature versioningdependency managementWeb UI | Available | Data remains in end-user's cloud accountSSOData encryption at restACL & RBAC | |
Iguazio | www.iguazio.com | https://youtu.be/BzQQ1X4LgcQ | Open-sourceFully managed cloud serviceSelf-managed commercial | AWSGCPAzureOn-prem | Available | Available | Feature versioningWeb UIdependency management | Available | Data remains in end-user's cloud accountACL & RBAC | mlops.community |
Kascada | kaskada.com | N/A | ||||||||
Molecula | www.molecula.com | |||||||||
RasgoML | www.rasgoml.com | https://youtu.be/CqjdZLjfg4w | Open-sourceFully managed cloud service | AWSGCPAzureOn-prem | Available | Not available | Feature versioningdependency managementWeb UI | Available | Data remains in end-user's cloud account | mlops.community |
Continual | continual.ai | |||||||||
Scribble data | www.scribbledata.io | https://youtu.be/hdV-InVETxE | Self-managed commercialFully managed cloud service | AWSGCPOn-prem | Available | Available | Feature versioningdependency managementWeb UI | Available | Data remains in end-user's cloud accountACL & RBACSSOData encryption at rest | mlops.community |
Databricks | mlops.community | |||||||||
AWS | Fully managed cloud service | mlops.community | ||||||||
Bytehub | www.bytehub.ai | https://youtu.be/ucAlzaJoqeU | Open-sourceFully managed cloud service |
Feature Store for ML evaluation and comparison
Compare providers Criteria 1 First, you need to assess whether the product's commercial characteristics meet your needs. We recommend evaluating the following commercial criteria: Delivery Model: Open source or managed service? Standalone feature store or part of a broader ML platform? Is the product available on-premises and / or in your public cloud?
mlops.community
