On this page
Seamlessly integrate Neptune's advanced experiment tracking features into your ZenML workflows to optimize your machine learning experimentation process. Leverage Neptune's intuitive UI to log, visualize, and compare pipeline runs, making it easier to identify the best performing models and iterate faster.
Features with ZenML
- Effortlessly track and visualize ZenML pipeline runs in Neptune
- Log models, parameters, metrics, and artifacts from pipeline steps
- Seamlessly transition from experimentation to production workflows
- Share pipeline results with team members and stakeholders via Neptune
- Leverage Neptune as a model registry for production-ready models
Main Features
- Interactive experiment tracking and visualization
- Comprehensive logging of metrics, parameters, artifacts, and more
- Collaborative workspace for sharing results with team members
- Flexible and customizable experiment management
- Integrates with popular ML frameworks and libraries
How to use ZenML with Neptune
import numpy as np
from zenml import step
from zenml.integrations.neptune.experiment_trackers.run_state import (
get_neptune_run,
)
from zenml.integrations.neptune.flavors import NeptuneExperimentTrackerSettings
neptune_settings = NeptuneExperimentTrackerSettings(tags={"classifier", "mnist"})
@step(
experiment_tracker="<NEPTUNE_TRACKER_STACK_COMPONENT_NAME>",
settings={
"experiment_tracker": neptune_settings
}
)
def training_step(
x_test: np.ndarray,
y_test: np.ndarray,
model,
) -> float:
"""Log metadata to Neptune run"""
neptune_run = get_neptune_run()
neptune_run["metrics"] = ...
...
