Database dump helps you restore data on your models, databases, and inferences when you start a new Docker* container. To get the current database state, access Settings in the upper-right corner of the interface and click Download Database:
The database is downloaded as a workbench_db_dump_<timestamp>.tar.gz
archive both to your Downloads folder and to a mounted folder, if any.
If you have not downloaded the dump in the previous session, only a clean start is available. Follow steps in the Installation Guide.
If you restore the DL Workbench state only with the database dump, the DL Workbench marks models, datasets, and projects as Read only.
Read-only artifacts on the Create Configuration page:
Read-only project on the Configurations page:
By default, models and datasets are saved to the /home/openvino/.workbench
directory in the Docker* container. If you launch the DL Workbench Docker container with your local folder mounted to /home/openvino/.workbench
, models and datasets are saved to your local folder as well.
To restore the DL Workbench state with models and datasets, make sure you have them in your local folder that you will mount to the /home/openvino/.workbench
directory in the Docker container when restoring the database state.
The structure of a mounted folder should look like this:
To restore the database state, take the two steps when running the Docker image:
workbench_db_dump.tar.gz
archive to the /home/openvino/.workbench
directory in the Docker container.DATABASE_DUMP_FILE
, which represents the dump file name.The examples below show how to mount the local folder ~/.workbench
or /C/workbench
and provide the environmental variable DATABASE_DUMP_FILE
while running a Docker container with the DL Workbench in the detached mode on CPU devices. To learn how to run the application on different targets or in a different mode, see Install DL Workbench from Docker Hub.
start_workbench.sh
script as described in the Install DL Workbench from Docker Hub* on Linux* OS section of Install DL Workbench from Docker Hub.NOTE: Replace
<timestamp>
with the digits in the name of your dump file.
For details about other parameters, see the Parameter Explanation section of Install DL Workbench from Docker Hub.
The following pop-up message indicates the state is successfully restored: