原书出处
Reliable Machine Learning
10 Continuous ML
But in addition to giving what we hope are useful mitigation strategies for creating robust and reliable systems even in the face of the difficulties, we would like to make a broader recommendation: All production ML systems should be treated as continuous ML systems. We should think about every model that is a key part of a production system as one that is trained continuously, even if it is not actually updated on new data e
原书定位详情
来源:a19c7933-4eda-4176-aebd-30fe7e1ad292
正文位置:extracted/Reliable Machine Learning/_epub-src/OEBPS/ch10.xhtml
EPUB SHA-256:1201425671b4d9d06d56459bced6f75ec74a7aac380ea24f5b573c80fee7be65