Machine Learning Orchestration
DBND an open source framework for building and tracking data pipelines. DBND is used for processes ranging from data ingestion, preparation, machine learning model training and production.
DBND includes a Python library, set of APIs, and CLI that enables you to collect metadata from your workflows, create a system of record for runs, and easily orchestrate complex processes.
DBND simplifies the process of building and running data pipelines from dbnd import task
from dbnd import task
@task
def say_hello(name: str = "databand.ai") -> str:
value = "Hello %s!" % name
return value
And makes it easy to track your critical pipeline metadata
from dbnd import log_metric, log_dataframe
log_dataframe("my_dataset", my_dataset)
log_metric("r2", r2)
See our documentation with examples and quickstart guides to get up and running with DBND.
For using DBND, we recommend that you work with a virtual environment like Virtualenv or Conda. Update to the latest and greatest:
pip install dbnd
If you would like access to our latest features, or have any questions, feedback, or contributions we would love to here from you! Get in touch through contact@databand.ai