Monitoring High-Performance Machine Learning Models with RAPIDS and whylogs

RAPIDS and cuML are used by organizations to run machine learning experiments faster on larger datasets. However, monitoring high-performance ML models can be challenging due to the potential for model failures caused by bad data. To address this issue, WhyLabs developed an open-source library called whylogs, which enables a lightweight data monitoring layer throughout the MLOps pipeline at scale. By integrating whylogs with RAPIDS and cuML, users can increase their speed while having a unified framework to detect data quality issues and data drift regardless of training data size. The infrastructure-agnostic approach means that anyone using RAPIDS can easily plug whylogs into their workflows with just a few lines of code.