Model Fields

For a more condensed summary this information is available in table view

The specification for Model has these fields:

name

Required-string
Name of the model

version

Optional-string
Version of the model, defined by the provider.

Example

"0.1"

license

Optional-string
License of the model. Type tbd in SAD group

provider_contacts

Optional-string
Dictionary with keys being phone, email, address - Type takes into account co-simulation models, for which solver info is mandatory (next slide)

description

Optional-string
Model tag description Additional info about model files (e.g. versioning, scope, i.e. what is the model used for, e.g. simulation, control, etc.) - Human readable marketplace description

Example

This is a brilliant model for doing complex things.

auth_tool

Optional-string
Authoring Tool used to create the model

type

Optional-string
ML, LCA, 3D FEM and CFD, System simulation, discrete event simulation, co-simulation (it couples a model with specific algorithms) - The original term Path was proposed to be changed (SAD)

fidelity

Optional-integer
Error of the model’s prediction

repository_uri

Required-string
Where the model file is stored (usually the DigitBrain certified external model repository). The path and model filename are not provided via this field.

Example

"https://www.myrepo.com"

path

Required-string
Path to the model file in the specified repository, not including the filename itself.

Example

"input/models"

filename

Required-string
Name of the model file at the given path within the given repo, with a file extension if it exists.

Example

"mymodel.pb"

state_depend

Optional-boolean
stateful -> 1, stateless -> 0

inslots

Optional-Slots[]
List of objects with values on which the model is evaluated (or parameters that might set before or during the evaluation) - Inputs and parameters together, although if needed we can still differentiate them internally in key field for each specific model.

Example

[
              { 
                "Key": "MY_INSLO_A",
                "Name": "ModelABin",
                "Dimensions": 112,
                "Is-continuous": true,
                "Units": {
                    "Unit": "Pa",
                    "Exponent": [1, -2, 0],
                    "Offset": 442,
                    "Scale": 10^-3
                },
                "Default-value": 55,
                "Ranges": [1,120]
              }
            ]  

outslots

Optional-Slots[]
List of objects with values that the model approximates

Example

[
              { 
                "Key": "MY_OUTPUT_A",
                "Name": "ModelABout",
                "Dimensions": 112,
                "Is-continuous": true,
                "Units": {
                    "Unit": "Pa",
                    "Exponent": [1, -2, 0],
                    "Offset": 442,
                    "Scale": 10^-3
                },
                "Default-value": 55,
                "Ranges": [1,120]
              }
            ]  

os_arch

Optional-string
OS architecture type (e.g. x86_64)

os_type

Optional-string
OS type (e.g. Windows, Linux)

os_distribution

Optional-string
OS distributun (e.g. Ubuntu, Fedora)

recommended_number_of_gpu_cores

Optional-integer
Recommended number of GPU cores

minimum_number_of_gpu_cores

Optional-integer
Minimum required number of GPU cores

recommended_gpu_ram

Optional-integer
Recommended GPU memory

minimum_gpu_ram

Optional-integer
Minimum required GPU memory

recommended_ram

Optional-integer
Recommended Memory

minimum_ram

Optional-integer
Minimum required memory

recommended_cpus

Optional-integer
Recommended number of CPU cores

minimum_cpus

Optional-integer
Minimum required number of CPU cores

required_disk_space

Optional-integer
Required amount of disk space in GB