Why Does the .modelarts Directory Automatically Appear in a Notebook Instance?
When you create or start a notebook instance, the Notebook function automatically generates configuration files and scripts in the $WORK_DIR/.modelarts directory to support functions such as pipeline, performance analysis plugin, AI Agent plugin, and Code Server. This section describes the structure of the .modelarts configuration directory and the usage of each file.
$WORK_DIR is the working directory of the notebook instance, and $INSTANCE_ID is the instance ID. Replace them with the actual values.
Precautions
Do not manually delete files in the .modelarts directory. Otherwise, related functions may be unavailable.
JupyterLab configuration directory
The following files are automatically generated when you use JupyterLab functions:
| Path | Generation Time | Description |
|---|---|---|
| $WORK_DIR/.modelarts/credentials | When a notebook instance is started | Pipeline template configuration file |
| $WORK_DIR/.modelarts/advisor | When the Performance Advisor plugin is used | Configuration path of the performance analysis plugin |
| $WORK_DIR/.modelarts/{Plugin name abbreviation}_check_status.sh | When the AI Agent plugin is used | Detection script of the AI Agent plugin |
| $WORK_DIR/.modelarts/install_{Plugin name abbreviation}.sh | When the AI Agent plugin is used | Installation script of the AI Agent plugin |
Code Server configuration directories
The following directories are automatically generated after you select Code Server when creating a notebook instance.
| Path | Description |
|---|---|
| $WORK_DIR/.modelarts/$INSTANCE_ID/code-server | Code Server user configuration directory |
| $WORK_DIR/.modelarts/$INSTANCE_ID/code-server-extensions | Code Server plugin directory |
What is your overall rating for this page?
Thank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbot