Comparison with MlFlow

What has our experiment tracking solved for?

  • Role Based Access Control- With MLFoundry we solve for one of the most major pain points of MLFlow which is allowing users to track their ML models and share at the level they feel comfortable.
  • Secure by design- Hosted collaborative dashboards without needing to share cloud credentials with every developer.
  • First-class support for non-scalar metrics - You don’t have to log your histograms, confusion-matrix and other non-scalar metrics as generic artifacts which make them interactive.
  • First-class dataset logging support - Logging and versioning dataset which was used to train the model and visualizing data distributions.
  • System Metric Tracking - Monitor and optimize your cloud-cost by keeping track of resource consumption like CPU, GPU & Memory.
  • Fast and Intuitive UI experience - Seamless UI which allows you to get a birds eye view of all your projects and dive deeper into every run to be able to visualize data, code, metrics and artifacts all in one place.