Skip to content

Class: Adapter

Adapter-based methods add extra trainable parameters after the attention and fully-connected layers of a frozen pretrained model to reduce memory-usage and speed up training. The adapters are typically small but demonstrate comparable performance to a fully finetuned model and enable training larger models with fewer resources. (https://huggingface.co/docs/peft/en/conceptual_guides/adapter)

URI: nexus:Adapter

 classDiagram
    class Adapter
    click Adapter href "../Adapter/"
      LargeLanguageModel <|-- Adapter
        click LargeLanguageModel href "../LargeLanguageModel/"
      Entry <|-- Adapter
        click Entry href "../Entry/"

      Adapter : adaptsModel





        Adapter --> "0..1" LargeLanguageModel : adaptsModel
        click LargeLanguageModel href "../LargeLanguageModel/"



      Adapter : architecture

      Adapter : broad_mappings





        Adapter --> "*" Any : broad_mappings
        click Any href "../Any/"



      Adapter : carbon_emitted

      Adapter : close_mappings





        Adapter --> "*" Any : close_mappings
        click Any href "../Any/"



      Adapter : contextWindowSize

      Adapter : dateCreated

      Adapter : dateModified

      Adapter : description

      Adapter : exact_mappings





        Adapter --> "*" Any : exact_mappings
        click Any href "../Any/"



      Adapter : fine_tuning

      Adapter : gpu_hours

      Adapter : hasAdapterType





        Adapter --> "0..1" AdapterType : hasAdapterType
        click AdapterType href "../AdapterType/"



      Adapter : hasCapability





        Adapter --> "*" Capability : hasCapability
        click Capability href "../Capability/"



      Adapter : hasDocumentation





        Adapter --> "*" Documentation : hasDocumentation
        click Documentation href "../Documentation/"



      Adapter : hasEvaluation





        Adapter --> "*" AiEvalResult : hasEvaluation
        click AiEvalResult href "../AiEvalResult/"



      Adapter : hasInputModality





        Adapter --> "*" Modality : hasInputModality
        click Modality href "../Modality/"



      Adapter : hasLicense





        Adapter --> "0..1" License : hasLicense
        click License href "../License/"



      Adapter : hasModelCard

      Adapter : hasOutputModality





        Adapter --> "*" Modality : hasOutputModality
        click Modality href "../Modality/"



      Adapter : hasRelatedRisk





        Adapter --> "*" Risk : hasRelatedRisk
        click Risk href "../Risk/"



      Adapter : hasRiskControl





        Adapter --> "*" RiskControl : hasRiskControl
        click RiskControl href "../RiskControl/"



      Adapter : hasTrainingData





        Adapter --> "*" Dataset : hasTrainingData
        click Dataset href "../Dataset/"



      Adapter : id

      Adapter : implementedByAdapter





        Adapter --> "*" Adapter : implementedByAdapter
        click Adapter href "../Adapter/"



      Adapter : implementsCapability





        Adapter --> "*" Capability : implementsCapability
        click Capability href "../Capability/"



      Adapter : isDefinedByTaxonomy





        Adapter --> "0..1" Taxonomy : isDefinedByTaxonomy
        click Taxonomy href "../Taxonomy/"



      Adapter : isDefinedByVocabulary





        Adapter --> "0..1" Vocabulary : isDefinedByVocabulary
        click Vocabulary href "../Vocabulary/"



      Adapter : isPartOf





        Adapter --> "0..1" LargeLanguageModelFamily : isPartOf
        click LargeLanguageModelFamily href "../LargeLanguageModelFamily/"



      Adapter : isProvidedBy





        Adapter --> "0..1" AiProvider : isProvidedBy
        click AiProvider href "../AiProvider/"



      Adapter : name

      Adapter : narrow_mappings





        Adapter --> "*" Any : narrow_mappings
        click Any href "../Any/"



      Adapter : numParameters

      Adapter : numTrainingTokens

      Adapter : performsTask





        Adapter --> "*" AiTask : performsTask
        click AiTask href "../AiTask/"



      Adapter : power_consumption_w

      Adapter : producer





        Adapter --> "0..1" Organization : producer
        click Organization href "../Organization/"



      Adapter : related_mappings





        Adapter --> "*" Any : related_mappings
        click Any href "../Any/"



      Adapter : requiredByTask





        Adapter --> "*" AiTask : requiredByTask
        click AiTask href "../AiTask/"



      Adapter : requiresCapability





        Adapter --> "*" Capability : requiresCapability
        click Capability href "../Capability/"



      Adapter : supported_languages

      Adapter : type

      Adapter : url

Inheritance

Slots

Name Cardinality and Range Description Inheritance
hasAdapterType 0..1
AdapterType		 The Adapter type, for example: LORA, ALORA, X-LORA direct
isDefinedByVocabulary 0..1
Vocabulary		 A relationship where a term or a term group is defined by a vocabulary direct
hasDocumentation *
Documentation		 Indicates documentation associated with an entity direct
hasLicense 0..1
License		 Indicates licenses associated with a resource direct
hasRelatedRisk *
Risk or 
RiskConcept or 
Term		 A relationship where an entity relates to a risk direct
adaptsModel 0..1
LargeLanguageModel		 The LargeLanguageModel being adapted direct
implementsCapability *
Capability		 Indicates that this adapter implements a specific capability direct
hasCapability *
Capability		 Indicates the technical capabilities this entry possesses direct
requiresCapability *
Capability		 Indicates that this entry requires a specific capability direct
numParameters 0..1
Integer		 A property indicating the number of parameters in a LLM LargeLanguageModel
numTrainingTokens 0..1
Integer		 The number of tokens a AI model was trained on LargeLanguageModel
contextWindowSize 0..1
Integer		 The total length, in bytes, of an AI model's context window LargeLanguageModel
hasInputModality *
Modality		 A relationship indicating the input modalities supported by an AI component LargeLanguageModel
hasOutputModality *
Modality		 A relationship indicating the output modalities supported by an AI component LargeLanguageModel
hasTrainingData *
Dataset		 A relationship indicating the datasets an AI model was trained on LargeLanguageModel
fine_tuning 0..1
String		 A description of the fine-tuning mechanism(s) applied to a model LargeLanguageModel
supported_languages *
String		 A list of languages, expressed as ISO two letter codes LargeLanguageModel
isPartOf 0..1
LargeLanguageModelFamily		 Annotation that a Large Language model is part of a family of models LargeLanguageModel, Entry
isDefinedByTaxonomy 0..1
Taxonomy		 A relationship where a concept or a concept group is defined by a taxonomy Entry
requiredByTask *
AiTask		 Indicates that this entry is required to perform a specific AI task Entry
implementedByAdapter *
Adapter		 Indicates that this capability is implemented by a specific adapter Entry
type 0..1
String		 Entry
id 1
String		 A unique identifier to this instance of the model element Entity
name 0..1
String		 A text name of this instance Entity
description 0..1
String		 The description of an entity Entity
url 0..1
Uri		 An optional URL associated with this instance Entity
dateCreated 0..1
Date		 The date on which the entity was created Entity
dateModified 0..1
Date		 The date on which the entity was most recently modified Entity
exact_mappings *
Any		 The property is used to link two concepts, indicating a high degree of confid... Entity
close_mappings *
Any		 The property is used to link two concepts that are sufficiently similar that ... Entity
related_mappings *
Any		 The property skos:relatedMatch is used to state an associative mapping link b... Entity
narrow_mappings *
Any		 The property is used to state a hierarchical mapping link between two concept... Entity
broad_mappings *
Any		 The property is used to state a hierarchical mapping link between two concept... Entity
hasEvaluation *
AiEvalResult		 A relationship indicating that an entity has an AI evaluation result AiModel
architecture 0..1
String		 A description of the architecture of an AI such as 'Decoder-only' AiModel
gpu_hours 0..1
Integer		 GPU consumption in terms of hours AiModel
power_consumption_w 0..1
Integer		 power consumption in Watts AiModel
carbon_emitted 0..1
Float		 The number of tons of carbon dioxide equivalent that are emitted during train... AiModel
hasRiskControl *
RiskControl		 Indicates the control measures associated with a system or component to modif... AiModel
producer 0..1
Organization		 A relationship to the Organization instance which produces this instance BaseAi
hasModelCard *
String		 A relationship to model card references BaseAi
performsTask *
AiTask		 relationship indicating the AI tasks an AI model can perform BaseAi
isProvidedBy 0..1
AiProvider		 A relationship indicating the AI model has been provided by an AI model provi... BaseAi

Usages

used by used in type used
Container adapters range Adapter
Entry implementedByAdapter range Adapter
Term implementedByAdapter range Adapter
Principle implementedByAdapter range Adapter
Risk implementedByAdapter range Adapter
AiTask implementedByAdapter range Adapter
Capability implementedByAdapter range Adapter
Adapter implementedByAdapter range Adapter
LLMIntrinsic hasAdapter range Adapter
LLMIntrinsic implementedByAdapter range Adapter

Identifier and Mapping Information

Schema Source

  • from schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology

Mappings

Mapping Type Mapped Value
self nexus:Adapter
native nexus:Adapter

LinkML Source

Direct

name: Adapter
description: Adapter-based methods add extra trainable parameters after the attention
  and fully-connected layers of a frozen pretrained model to reduce memory-usage and
  speed up training. The adapters are typically small but demonstrate comparable performance
  to a fully finetuned model and enable training larger models with fewer resources.
  (https://huggingface.co/docs/peft/en/conceptual_guides/adapter)
from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
is_a: Entry
mixins:
- LargeLanguageModel
slots:
- hasAdapterType
- isDefinedByVocabulary
- hasDocumentation
- hasLicense
- hasRelatedRisk
- adaptsModel
- implementsCapability
- hasCapability
- requiresCapability

Induced

name: Adapter
description: Adapter-based methods add extra trainable parameters after the attention
  and fully-connected layers of a frozen pretrained model to reduce memory-usage and
  speed up training. The adapters are typically small but demonstrate comparable performance
  to a fully finetuned model and enable training larger models with fewer resources.
  (https://huggingface.co/docs/peft/en/conceptual_guides/adapter)
from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
is_a: Entry
mixins:
- LargeLanguageModel
attributes:
  hasAdapterType:
    name: hasAdapterType
    description: 'The Adapter type, for example: LORA, ALORA, X-LORA'
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: hasAdapterType
    owner: Adapter
    domain_of:
    - Adapter
    range: AdapterType
  isDefinedByVocabulary:
    name: isDefinedByVocabulary
    description: A relationship where a term or a term group is defined by a vocabulary
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: schema:isPartOf
    alias: isDefinedByVocabulary
    owner: Adapter
    domain_of:
    - Entry
    - Term
    - Adapter
    - LLMIntrinsic
    range: Vocabulary
  hasDocumentation:
    name: hasDocumentation
    description: Indicates documentation associated with an entity.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: airo:hasDocumentation
    alias: hasDocumentation
    owner: Adapter
    domain_of:
    - Dataset
    - Vocabulary
    - Taxonomy
    - Concept
    - Group
    - Entry
    - Term
    - Principle
    - RiskTaxonomy
    - Action
    - BaseAi
    - LargeLanguageModelFamily
    - AiEval
    - BenchmarkMetadataCard
    - Adapter
    - LLMIntrinsic
    range: Documentation
    multivalued: true
    inlined: false
  hasLicense:
    name: hasLicense
    description: Indicates licenses associated with a resource
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: airo:hasLicense
    alias: hasLicense
    owner: Adapter
    domain_of:
    - Dataset
    - Documentation
    - Vocabulary
    - Taxonomy
    - RiskTaxonomy
    - BaseAi
    - AiEval
    - BenchmarkMetadataCard
    - Adapter
    range: License
  hasRelatedRisk:
    name: hasRelatedRisk
    description: A relationship where an entity relates to a risk
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    domain: Any
    alias: hasRelatedRisk
    owner: Adapter
    domain_of:
    - Term
    - LLMQuestionPolicy
    - Action
    - AiEval
    - BenchmarkMetadataCard
    - Adapter
    - LLMIntrinsic
    range: Risk
    multivalued: true
    inlined: false
    any_of:
    - range: RiskConcept
    - range: Term
  adaptsModel:
    name: adaptsModel
    description: The LargeLanguageModel being adapted
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: adaptsModel
    owner: Adapter
    domain_of:
    - Adapter
    range: LargeLanguageModel
  implementsCapability:
    name: implementsCapability
    description: Indicates that this adapter implements a specific capability
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    domain: Any
    alias: implementsCapability
    owner: Adapter
    domain_of:
    - Adapter
    inverse: implementedByAdapter
    range: Capability
    multivalued: true
    inlined: false
  hasCapability:
    name: hasCapability
    description: 'Indicates the technical capabilities this entry possesses.

      '
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: tech:hasCapability
    alias: hasCapability
    owner: Adapter
    domain_of:
    - AiSystem
    - Adapter
    - LLMIntrinsic
    range: Capability
    multivalued: true
    inlined: false
  requiresCapability:
    name: requiresCapability
    description: Indicates that this entry requires a specific capability
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    domain: Any
    alias: requiresCapability
    owner: Adapter
    domain_of:
    - Entry
    - LargeLanguageModel
    - AiTask
    - Adapter
    inverse: requiredByTask
    range: Capability
    multivalued: true
    inlined: false
  numParameters:
    name: numParameters
    description: A property indicating the number of parameters in a LLM.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: numParameters
    owner: Adapter
    domain_of:
    - LargeLanguageModel
    range: integer
    minimum_value: 0
  numTrainingTokens:
    name: numTrainingTokens
    description: The number of tokens a AI model was trained on.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: numTrainingTokens
    owner: Adapter
    domain_of:
    - LargeLanguageModel
    range: integer
    minimum_value: 0
  contextWindowSize:
    name: contextWindowSize
    description: The total length, in bytes, of an AI model's context window.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: contextWindowSize
    owner: Adapter
    domain_of:
    - LargeLanguageModel
    range: integer
    minimum_value: 0
  hasInputModality:
    name: hasInputModality
    description: A relationship indicating the input modalities supported by an AI
      component. Examples include text, image, video.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: hasInputModality
    owner: Adapter
    domain_of:
    - LargeLanguageModel
    range: Modality
    multivalued: true
    inlined: false
  hasOutputModality:
    name: hasOutputModality
    description: A relationship indicating the output modalities supported by an AI
      component. Examples include text, image, video.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: hasOutputModality
    owner: Adapter
    domain_of:
    - LargeLanguageModel
    range: Modality
    multivalued: true
    inlined: false
  hasTrainingData:
    name: hasTrainingData
    description: A relationship indicating the datasets an AI model was trained on.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: airo:hasTrainingData
    alias: hasTrainingData
    owner: Adapter
    domain_of:
    - LargeLanguageModel
    range: Dataset
    multivalued: true
    inlined: false
  fine_tuning:
    name: fine_tuning
    description: A description of the fine-tuning mechanism(s) applied to a model.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: fine_tuning
    owner: Adapter
    domain_of:
    - LargeLanguageModel
    range: string
  supported_languages:
    name: supported_languages
    description: A list of languages, expressed as ISO two letter codes. For example,
      'jp, fr, en, de'
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: supported_languages
    owner: Adapter
    domain_of:
    - LargeLanguageModel
    range: string
    multivalued: true
    inlined: true
    inlined_as_list: true
  isPartOf:
    name: isPartOf
    description: Annotation that a Large Language model is part of a family of models
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: schema:isPartOf
    alias: isPartOf
    owner: Adapter
    domain_of:
    - Entry
    - Risk
    - LargeLanguageModel
    - CapabilityGroup
    - Stakeholder
    range: LargeLanguageModelFamily
  isDefinedByTaxonomy:
    name: isDefinedByTaxonomy
    description: A relationship where a concept or a concept group is defined by a
      taxonomy
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: schema:isPartOf
    alias: isDefinedByTaxonomy
    owner: Adapter
    domain_of:
    - Concept
    - Control
    - Group
    - Entry
    - Policy
    - RiskGroup
    - Risk
    - RiskControl
    - Action
    - RiskIncident
    - CapabilityGroup
    - StakeholderGroup
    - Stakeholder
    range: Taxonomy
  requiredByTask:
    name: requiredByTask
    description: Indicates that this entry is required to perform a specific AI task.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: requiredByTask
    owner: Adapter
    domain_of:
    - Entry
    - Capability
    inverse: requiresCapability
    range: AiTask
    multivalued: true
    inlined: false
  implementedByAdapter:
    name: implementedByAdapter
    description: 'Indicates that this capability is implemented by a specific adapter.
      This relationship distinguishes the abstract capability (what can be done) from
      the technical implementation mechanism (how it is added/extended via adapters).

      '
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    domain: Any
    alias: implementedByAdapter
    owner: Adapter
    domain_of:
    - Entry
    - Capability
    inverse: implementsCapability
    range: Adapter
    multivalued: true
    inlined: false
  type:
    name: type
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/common
    designates_type: true
    alias: type
    owner: Adapter
    domain_of:
    - Vocabulary
    - Taxonomy
    - Concept
    - Control
    - Group
    - Entry
    - Policy
    range: string
  id:
    name: id
    description: A unique identifier to this instance of the model element. Example
      identifiers include UUID, URI, URN, etc.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: schema:identifier
    identifier: true
    alias: id
    owner: Adapter
    domain_of:
    - Entity
    range: string
    required: true
  name:
    name: name
    description: A text name of this instance.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: schema:name
    alias: name
    owner: Adapter
    domain_of:
    - Entity
    - BenchmarkMetadataCard
    range: string
  description:
    name: description
    description: The description of an entity
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: schema:description
    alias: description
    owner: Adapter
    domain_of:
    - Entity
    range: string
  url:
    name: url
    description: An optional URL associated with this instance.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: schema:url
    alias: url
    owner: Adapter
    domain_of:
    - Entity
    range: uri
  dateCreated:
    name: dateCreated
    description: The date on which the entity was created.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: schema:dateCreated
    alias: dateCreated
    owner: Adapter
    domain_of:
    - Entity
    range: date
    required: false
  dateModified:
    name: dateModified
    description: The date on which the entity was most recently modified.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: schema:dateModified
    alias: dateModified
    owner: Adapter
    domain_of:
    - Entity
    range: date
    required: false
  exact_mappings:
    name: exact_mappings
    description: The property is used to link two concepts, indicating a high degree
      of confidence that the concepts can be used interchangeably across a wide range
      of information retrieval applications
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: skos:exactMatch
    alias: exact_mappings
    owner: Adapter
    domain_of:
    - Entity
    range: Any
    multivalued: true
    inlined: false
  close_mappings:
    name: close_mappings
    description: The property is used to link two concepts that are sufficiently similar
      that they can be used interchangeably in some information retrieval applications.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: skos:closeMatch
    alias: close_mappings
    owner: Adapter
    domain_of:
    - Entity
    range: Any
    multivalued: true
    inlined: false
  related_mappings:
    name: related_mappings
    description: The property skos:relatedMatch is used to state an associative mapping
      link between two concepts.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: skos:relatedMatch
    alias: related_mappings
    owner: Adapter
    domain_of:
    - Entity
    range: Any
    multivalued: true
    inlined: false
  narrow_mappings:
    name: narrow_mappings
    description: The property is used to state a hierarchical mapping link between
      two concepts, indicating that the concept linked to, is a narrower concept than
      the originating concept.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: skos:narrowMatch
    alias: narrow_mappings
    owner: Adapter
    domain_of:
    - Entity
    range: Any
    multivalued: true
    inlined: false
  broad_mappings:
    name: broad_mappings
    description: The property is used to state a hierarchical mapping link between
      two concepts, indicating that the concept linked to, is a broader concept than
      the originating concept.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: skos:broadMatch
    alias: broad_mappings
    owner: Adapter
    domain_of:
    - Entity
    range: Any
    multivalued: true
    inlined: false
  hasEvaluation:
    name: hasEvaluation
    description: A relationship indicating that an entity has an AI evaluation result.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: dqv:hasQualityMeasurement
    alias: hasEvaluation
    owner: Adapter
    domain_of:
    - AiModel
    range: AiEvalResult
    multivalued: true
  architecture:
    name: architecture
    description: A description of the architecture of an AI such as 'Decoder-only'.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: architecture
    owner: Adapter
    domain_of:
    - AiModel
    range: string
  gpu_hours:
    name: gpu_hours
    description: GPU consumption in terms of hours
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: gpu_hours
    owner: Adapter
    domain_of:
    - AiModel
    range: integer
    minimum_value: 0
  power_consumption_w:
    name: power_consumption_w
    description: power consumption in Watts
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: power_consumption_w
    owner: Adapter
    domain_of:
    - AiModel
    range: integer
    minimum_value: 0
  carbon_emitted:
    name: carbon_emitted
    description: The number of tons of carbon dioxide equivalent that are emitted
      during training
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: carbon_emitted
    owner: Adapter
    domain_of:
    - AiModel
    range: float
    minimum_value: 0
    unit:
      symbol: t CO2-eq
      descriptive_name: tons of CO2 equivalent
  hasRiskControl:
    name: hasRiskControl
    description: Indicates the control measures associated with a system or component
      to modify risks.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: airo:hasRiskControl
    alias: hasRiskControl
    owner: Adapter
    domain_of:
    - AiModel
    range: RiskControl
    multivalued: true
  producer:
    name: producer
    description: A relationship to the Organization instance which produces this instance.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: producer
    owner: Adapter
    domain_of:
    - BaseAi
    range: Organization
  hasModelCard:
    name: hasModelCard
    description: A relationship to model card references.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: hasModelCard
    owner: Adapter
    domain_of:
    - BaseAi
    range: string
    multivalued: true
    inlined: true
    inlined_as_list: true
  performsTask:
    name: performsTask
    description: relationship indicating the AI tasks an AI model can perform.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    alias: performsTask
    owner: Adapter
    domain_of:
    - BaseAi
    range: AiTask
    multivalued: true
    inlined: false
  isProvidedBy:
    name: isProvidedBy
    description: A relationship indicating the AI model has been provided by an AI
      model provider.
    from_schema: https://ibm.github.io/ai-atlas-nexus/ontology/ai-risk-ontology
    rank: 1000
    slot_uri: airo:isProvidedBy
    alias: isProvidedBy
    owner: Adapter
    domain_of:
    - BaseAi
    range: AiProvider