Datmo: Tracking, Versioning, and Environment handling tool for Machine Learning Models
Datmo's Community Edition is a workflow tool consisting of both a command line interface (CLI) and website, where developers can track, share, and deploy data, machine learning, and artificial intelligence models.
Who Uses Datmo?
If you work with data models and code, then Datmo is for you.
We know this because if you're working on a machine learning project, you're likely facing at least one of the following issues:
- having trouble in setting up libraries and frameworks
- cannot reproduce your work on different machines
- struggling to reproduce colleague's work due to dependencies
- no history of historical configurations to recreate previous work
- overwriting weights files for each run
- keeping track of configurations used and metrics in a spreadsheet but still losing track of your work
Datmo is here to help. This documentation is for developers who want to leverage the Datmo CLI. The CLI has the following features:
- replicate trained models completely without juggling multiple tools
- automatically versions these models as you change configurations, code, environments, metrics or files
- enables sharing of your versions with other community members and colleagues
- enables production-grade deployment of trained models
The Datmo CLI works in tandem with the web application. We are currently in open beta, you can sign up for access via the sign up page.
We are in an early-release Beta. Expect some adventures and we value your feedback and new feature requests. :)