Diagnosing Automatic Notebook Restarts During Feature Extraction

Summary

This Chinese-language support exchange addresses a quantitative research notebook that restarts automatically after two features are added and feature extraction begins. The user reports that the visible CPU and memory figures have not reached their displayed limits, but the excerpt does not include diagnostic logs or a confirmed root-cause investigation.

The replies identify memory exhaustion as the likely explanation and recommend reducing the data scale by using fewer samples or features. One suggested troubleshooting step is to shift the start date of the training window by several days, which may clear or reset cached data. These are practical debugging suggestions rather than a general description of the platform’s restart mechanism. The excerpt gives no evidence that either step resolved this particular incident, and the visible resource readings alone do not establish why the process restarted.

Key ideas

• The reported restart occurs during feature extraction after new features are added.
• A reply attributes automatic kernel restarts to insufficient memory, even when displayed usage appears below a threshold.
• Reducing sample count or feature dimensionality is suggested as a way to lower resource demand.
• Changing the training window’s start date by several days is proposed as a way to reset cached data.
• The exchange does not provide logs or confirm which suggested step resolved the issue.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.