Neptune does not come with hardware monitoring built in.
However, you can use log_metrics() to log hardware consumption collected using external libraries like psutil, pynvml, etc.
We also provide a standalone Class that can be imported to your training scripts to automatically log common hardware metrics to Neptune.
The script is available in our GitHub repository, along with usage instructions, at scale-examples/utils/monitoring_tools/hardware_monitoring at main · neptune-ai/scale-examples.
You can also explore an example run showing the different hardware metrics and attributes captured and logged by this class at hardware_monitoring.
