Overview
Polyaxon automation is a set of tools for:
- Authoring DAGs
- Running hyperparameter tuning of ML models
- Running jobs and pipelines in parallel
- Running components on schedule
- Subscribing components to events
Features
- Lineage: Strongly typed inputs and outputs for a full provenance of parameters and data.
- Helpers: All workflows can be parameterized, cached by default, with pre-computed results, outputs, and artifacts.
- Massive scale: Fully distributed, several scheduling strategies, controlled concurrency, fault-tolerant, scale to multiple nodes, clusters, and hundreds of thousands of concurrent executions.
- Dynamic, customizable, and extensible: Use several runtimes in your graph of operations, experiments, jobs, services, distributed jobs, …