- 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
create-version
Description
Create a model version.
Archives programs in the model name directory. Upload to ABEJA Platform. If a large file is placed in the model name directory, archiving and uploading may take some time.
Notes
The version cannot be updated. Create a new version and delete the old version.
How to deploy any version
You can create multiple versions of a model, and you can specify and deploy any version by specifying any version during create-service
.
Synopsis
$ abeja model create-version [--help]
Usage: abeja model create-version [OPTIONS]
Create version & upload application
Options:
-m, --model_id TEXT Model identifier [required]
-v, --version TEXT Model version code [required]
-i, --image TEXT Base-image name [required]
-h, --handler TEXT Path to handler in the model archive.
-j, --job_definition_name TEXT Training job definition name
-t, --training_job_id TEXT Training Job id
--help Show this message and exit.
Options
-m
, --model_id
Specify model_id when create-model is created.(Requireed)
-v
, --version
Specify any version (string type). (e.g. 0.1 / v0.1 / etc… (Requireed)
-i
, --image
Specifies the image that runs the model. (Requireed) Click here for each image.
abeja-inc/all-gpu:18.10
abeja-inc/all-cpu:18.10
abeja-inc/all-gpu:19.04
abeja-inc/all-cpu:19.04
-h
, --handler
Specify a handler.
e.g.
main:handler
predict:handler
-j
, --job_definition_name
Specify the training job definition name.
-t
, --training_job_id
Specify the training job id.
Example
Create version
Command:
$ abeja model create-version --model_id 1215320870077 --version "0.1" --image "abeja-inc/all-cpu:18.10" --handler "main:handler"
Output:
{'version': '0.1', 'upload_url': 'https://abeja-model-api-src-xxxxxx 'version_id': 'ver-2345678901234'}