brain version assessment start
Syntax and usage
bonsai brain version assessment start
-b BRAIN_NAME | --brain-name BRAIN_NAME
-c CONCEPT_NAME | --concept-name CONCEPT_NAME
-f PATH_TO_CONFIG_FILE | --file PATH_TO_CONFIG_FILE
[-i NUMBER_OF_SIMULATORS | --instance-count NUMBER_OF_SIMULATORS ]
[ --simulator-package-name SIMULATOR_PACKAGE_NAME ]
[ -n ASSESSMENT_NAME | --name ASSESSMENT_NAME ]
[ --brain-version BRAIN_VERSION ]
[ --display-name ASSESSMENT_DISPLAY_NAME ]
[ -des ASSESSMENT_DESCRIPTION | --description ASSESSMENT_DESCRIPTION ]
[ --episode-iteration-limit EPISODE_ITERATION_LIMIT ]
[ --maximum-duration MAXIMUM_TIME_DURATION ]
[ -o OUTPUT_FORMAT | --output OUTPUT_FORMAT ]
[ --debug ]
bonsai brain version assessment start -h | --help
Start an assessment for the given brain version and concept, and log the simulator behavior.
Options
-b BRAIN_NAME
--brain-name BRAIN_NAME
Required, string.
Indicates the unique system name for the target brain. System names are case-insensitive and can contain letters, numbers, and dashes.
Example: --brain-name 'HiveMind'
-c CONCEPT_NAME
--concept-name CONCEPT_NAME
Required, string.
Indicates the target concept defined in the Inkling for the target brain version.
Example: --concept-name 'OptimizeHoney'
-f PATH_TO_CONFIG_FILE
--file PATH_TO_CONFIG_FILE
Required, string.
File path (relative or absolute) to a JSON assessment configuration file. The file contains a version number and a set of episode configurations. Each configuration determines an initial simulator state that a test episode is run on.
Example: --file 'LOCAL/PATH/TO/CONFIGURATION/file.json'
--simulator-package-name SIMULATOR_PACKAGE_NAME
Optional, string.
Indicates the unique system name for a managed simulator package. Package names are case-insensitive and can contain letters, numbers, and dashes.
Example: --simulator-package-name 'Apiary-HVAC'
-i NUMBER_OF_SIMULATOR_INSTANCES
--instance-count NUMBER_OF_SIMULATOR_INSTANCES
Optional, integer.
Indicates the number of simulator instances to be used during training with the simulator package. Overrides the configuration setting provided when the targeted simulator package was added to Bonsai.
Example: --instance-count 16
--name ASSESSMENT_NAME
Optional, string.
Indicates the unique assessment name. When name is not provided, Bonsai autogenerates a name. Users must specify the assessment name to update, stop, delete or see details about an assessment. If a separate display name is not provided, name is also used in the Bonsai UI.
Example: --name 'HiveMindEdgeCases'
--brain-version BRAIN_VERSION
Optional, integer.
Indicates the desired version for the target brain. Defaults to the latest version.
Example: --brain-version 2
--display-name DISPLAY_NAME
Optional, string.
Indicates the user-friendly name of the resource (brain, exported brain, ML model, etc.). Bonsai uses the display name in the UI and CLI output. Display names do not need to be unique. If you do not provide a display name, the brain name is used.
Example: --display-name 'The Hive Mind'
--description DESCRIPTION
--des DESCRIPTION
Optional, string.
Provides a user-friendly description for the resource (brain, assessment, simulator etc.) that explains its purpose.
Example: --description 'Main controller for apiary environment HVAC.'
--maximum-duration MAXIMUM_TIME_DURATION
Optional, string.
The maximum amount of time an assessment should run for. The default is 24
hours. The maximum allowed duration is 7 days. Duration format is
{DURATION}{UNITS}, where {DURATION} is the quantity of time and {UNITS}
is days (d), hours (h), or minutes (m). The default unit is hours.
Example: --maximum-duration '3d2h'
--episode-iteration-limit EPISODE_ITERATION_LIMIT
Optional, integer.
Maximum number of iterations per assessment episode, defaults to 1000.
Example: --episode-iteration-limit 250
-o OUTPUT_FORMAT
--output OUTPUT_FORMAT
Optional, enumerated string.
Sets the display format for responses received from the Bonsai server. The default format is tabular. Other supported options are noted below.
| Token | Output format |
|---|---|
json |
JSON |
Example: --output json
--debug
Optional.
Display verbose logging information in JSON. Includes request details from the CLI and response details from the Bonsai server.
Example: --debug
-h
--help
Optional.
Displays in-line help for the command and exit. Required options are not needed when using the help flag.
Example: --help
Examples
Run an assessment for the OptimizeHoney concept in the latest HiveMind brain version:
$ bonsai brain version assessment start \
--brain-name HiveMind \
--concept-name OptimizeHoney \
--file TestingEdgeCases.json \
--name TestingEdgeCases \
--simulator-package-name Apiary-HVAC \
--instance-count 16
Started assessment TestingEdgeCases for brain HiveMind version 5.