Yes, MLflow can automatically log the usage of system resources such as GPUs, CPUs, and memory. This can be done by configuring the appropriate tracking backend, such as using the MLflow Tracking Server or the MLflow file storage backend. The tracking backend will collect system resource usage data and store it alongside other experiment metadata. The resource usage data can then be viewed or queried using the MLflow UI or programmatic APIs. Additionally, MLflow can integrate with various performance profiling tools to provide more detailed system resource usage information.
Asked: 2023-06-27 12:21:28 +0000
Seen: 7 times
Last updated: Jun 27 '23