日本語
Currently under development, an experimental DataRobot API allows model administrators to download a Deployed model from DataRobot cluster X, upload it to DataRobot cluster Y then deploy and make requests from it.
About this Accelerator
Model Requirements
- Model to migrate must by using py3 or better
Supported Paths
These model migration paths are currently known to work out of the box
- DR8.x -> DR8.x
- DR8.x -> DR9
Spoiler
Attention: This notebook WILL NOT WORK using https://app.datarobot.com
Prerequisites
- This notebook must be able to write to the model directory, located in the same directory as this notebook
- For best results: run this notebook from the local file system
- Model to migrate must be a deployed model
- The Source and Destination users must have the "Enable Experimental API access" permission for process to work
- Must have an API key from the Source and Destination clusters
- Notebook must have connectivity to the Source and Destination clusters
- The DataRobot versions must be consistent with the Supported Paths
- In the case of a DataRobot 7 model, the migrator must have ssh access to the App Node of the cluster
- The Source and Destination DataRobot clusters must have the following in the config.yaml
app_configuration:
drenv_override:
WHITELIST_EXPERIMENTAL_API: true
EXPERIMENTAL_API_ACCESS: true
To get started, find the Source Settings section of the notebook and update the following:
source_host
source_apikey
deployment_id
destination_host
destination_apikey
