- Overview
- Getting Started Guide
- UserGuide
-
References
-
ABEJA Platform CLI
- CONFIG COMMAND
- DATALAKE COMMAND
- DATASET COMMAND
- TRAINING COMMAND
-
MODEL COMMAND
- check-endpoint-image
- check-endpoint-json
- create-deployment
- create-endpoint
- create-model
- create-service
- create-trigger
- create-version
- delete-deployment
- delete-endpoint
- delete-model
- delete-service
- delete-version
- describe-deployments
- describe-endpoints
- describe-models
- describe-service-logs
- describe-services
- describe-versions
- download-versions
- run-local
- run-local-server
- start-service
- stop-service
- submit-run
- update-endpoint
- startapp command
-
ABEJA Platform CLI
- FAQ
- Appendix
Model and Model Version
In general, A model means a constructed algorithm which has been trained by trained data. ABEJA Platform is facilitating the whole process of executing inference.
Model Version
ABEJA Platform allows version management for the created model. You also need to specify the version when creating a new model.
In a model, version needs to include requirements.txt defining the packages to be install, source code including the implementation of the model handler function, and the image of the model execution environment.
Model Handler Function
The model handler function is a python-implemented function called when the model is executed.
Add Library
Select an image from ABEJA Platform provided images which become the execution environment of mode. In each image, required deep learning libraries are pre-installed and can be used from the model source code. (Pre-installed libraries are described below). Required libraries can be installed directly by placing at requirements.txt at the time of creating the model version. Library mentioned will be installed at the time of creating the service.
Image
When creating a model version, you need to choose the environment in which the created version will be executed.