ALIYUN::PAI::ModelVersion类型用于创建模型版本。
语法
{
"Type": "ALIYUN::PAI::ModelVersion",
"Properties": {
"ModelId": String,
"Uri": String,
"ApprovalStatus": String,
"ExtraInfo": Map,
"FormatType": String,
"FrameworkType": String,
"InferenceSpec": Map,
"Labels": List,
"Metrics": Map,
"Options": String,
"SourceType": String,
"SourceId": String,
"TrainingSpec": Map,
"VersionDescription": String,
"VersionName": String
}
}属性
|
属性名称 |
类型 |
必须 |
允许更新 |
描述 |
约束 |
|
ModelId |
String |
是 |
否 |
模型ID。 |
无 |
|
Uri |
String |
是 |
否 |
模型版本URI，即模型存储位置。 |
可能的取值为： - 模型的HTTP(S)地址，例如：'https://myweb.com/mymodel.tar.gz '。 - 如果模型存储在OSS中，格式为'oss://./object '。关于endpoint配置，例如，参见OSS:// mybucket/'。 |
|
ApprovalStatus |
String |
否 |
是 |
审批状态。 |
取值：
|
|
ExtraInfo |
Map |
否 |
是 |
附加信息。 |
无 |
|
FormatType |
String |
否 |
否 |
模型格式， |
取值：
|
|
FrameworkType |
String |
否 |
否 |
模型框架。 |
取值：
|
|
InferenceSpec |
Map |
否 |
是 |
描述如何应用到下游推理服务。 |
例如，描述EAS的处理器和容器。示例：{ "processor": "tensorflow_gpu_1.12" } |
|
Labels |
List |
否 |
是 |
模型版本标签列表。 |
长度最大值为20。更多信息，请参考Labels属性。 |
|
Metrics |
Map |
否 |
是 |
模型指标。 |
序列化后的长度限制为8192。示例： { "Results": [{ "Dataset": { "DatasetId": "d-sdkjanksaklerhfd" }, "Metrics": { "cer": 0.175 } }, { "Dataset": { "Uri": "oss://xxxx/" }, "Metrics": { "cer": 0.172 } }] } |
|
Options |
String |
否 |
是 |
扩展字段。 |
这是一个JSON字符串。 |
|
SourceId |
String |
否 |
是 |
来源ID。 |
当来源类型为Custom时，此字段不受限制。 * 当来源为PAIFlow或TrainingService时，格式为 ''' region=,workspaceId=,kind=,id= ''' 其中： - region为阿里云地域ID。 - workspaceId表示工作空间ID。 - kind：类型。取值：PipelineRun（PAIFlow工作流）；ServiceJob（训练服务）。 - id：唯一标识符。 |
|
SourceType |
String |
否 |
是 |
模型来源类型。 |
取值：
|
|
TrainingSpec |
Map |
否 |
是 |
训练配置。 |
用于微调、增量训练的配置。 |
|
VersionDescription |
String |
否 |
是 |
模型版本描述。 |
无 |
|
VersionName |
String |
否 |
否 |
模型版本， |
对每个模型唯一。如果留空此参数，第一个版本默认为0.1.0。之后，次版本号依次递增1。例如，第二个版本号为0.2.0。版本号由主版本号、次版本号和阶段版本号组成，用点号（.）分隔。主版本号和次版本号为数字。阶段版本号以数字开头，可以包含数字、下划线和字母。例如，版本号为1.1.0或2.3.4_beta。 |
Labels语法
"Labels": [
{
"Value": String,
"Key": String
}
]Labels属性
|
属性名称 |
类型 |
必须 |
允许更新 |
描述 |
约束 |
|
Key |
String |
否 |
是 |
标签键。 |
无 |
|
Value |
String |
否 |
是 |
标签值。 |
无 |
返回值
Fn::GetAtt
-
Options：扩展字段。JsonString类型。
-
VersionDescription：模型版本描述。
-
SourceType：模型来源类型，可能的取值：
-
SourceId：来源ID。
-
ExtraInfo：其他信息。
-
Labels：模型版本标签列表。
-
TrainingSpec：训练配置。用于微调、增量训练的配置。
-
Uri：模型版本URI，即模型存储位置。
-
GmtModifiedTime：最后更新模型的UTC时间，格式为iso8601。
-
FormatType：模型格式，可能的取值：
-
ApprovalStatus：审批状态，取值如下：
-
Metrics：模型指标。
-
OwnerId：阿里云账号ID。
-
UserId：用户ID。
-
FrameworkType：模型框架，可能的取值：
-
InferenceSpec：描述如何应用到下游推理服务，例如EAS的处理器和容器。
-
VersionName：模型版本。
-
GmtCreateTime：创建模型的UTC时间，格式为ISO8601。
示例
ROSTemplateFormatVersion: '2015-09-01'
Parameters:
VersionDescription:
Type: String
Description:
en: Model version description.
Default: Null
Required: false
FormatType:
Type: String
Description:
en: |-
Model format, possible values:
- OfflineModel
- SavedModel
- Keras H5
- Frozen Pb
- Caffe Prototxt
- TorchScript
- XGBoost
- PMML
- AlinkModel
- ONNX.
AllowedValues:
- OfflineModel
- SavedModel
- Keras H5
- Frozen Pb
- Caffe Prototxt
- TorchScript
- XGBoost
- PMML
- AlinkModel
- ONNX
Default: Null
Required: false
SourceType:
Type: String
Description:
en: |-
Model source type, possible values:
- Custom: Custom.
- PAIFlow:PAI workflow.
- TrainingService:PAI training service.
AllowedValues:
- Custom
- PAIFlow
- TrainingService
Default: Null
Required: false
SourceId:
Type: String
Description:
en: |-
Source ID.
* When the source type is Custom, this field is not restricted.
* When the source is PAIFlow or TrainingService, the format is
'''
region=<region_id>,workspaceId=<workspace_id>,kind=<kind>,id=<id>
'''
Among them:
- region is the Alibaba Cloud region ID.
- workspaceId indicates the workspace ID.
- kind: is a type. Value: PipelineRun(PAIFlow workflow);ServiceJob (training service).
- id: is a unique identifier.
Default: Null
Required: false
ApprovalStatus:
Type: String
Description:
en: |-
Admission status, with values as follows:
- Pending: to be determined.
- Approved: Allow to go online.
- Rejected: Online is not allowed.
AllowedValues:
- Pending
- Approved
- Rejected
Default: Null
Required: false
ExtraInfo:
Description:
en: The additional information.
Required: false
Default: Null
Type: Json
FrameworkType:
Type: String
Description:
en: |-
Model framework, possible values:
- Pytorch
- XGBoost
- Keras
- Caffe
- Alink
- Xflow
- TensorFlow.
AllowedValues:
- Pytorch
- XGBoost
- Keras
- Caffe
- Alink
- Xflow
- TensorFlow
Default: Null
Required: false
InferenceSpec:
Description:
en: 'Describes how to apply to downstream inference services. For example, describe the processor and container of EAS. Example: { "processor": "tensorflow_gpu_1.12" }'
Required: false
Default: Null
Type: Json
Metrics:
Description:
en: |-
The metrics for the model. The length after serialization is limited to 8,192. Example:
{ "Results": [{ "Dataset": { "DatasetId": "d-sdkjanksaklerhfd" }, "Metrics": { "cer": 0.175 } }, { "Dataset": { "Uri": "oss://xxxx/" }, "Metrics": { "cer": 0.172 } }] }
Required: false
Default: Null
Type: Json
Labels:
Description:
en: List of model version labels.
Required: false
Default: Null
Type: Json
MaxLength: 20
AssociationProperty: List[Parameters]
AssociationPropertyMetadata:
Parameters:
Key:
Type: String
Description:
en: label key.
Default: Null
Required: false
Value:
Type: String
Description:
en: label value.
Default: Null
Required: false
VersionName:
Type: String
Description:
en: The model version, which is unique for each model. If you leave this parameter empty, the first version is 0.1.0 by default. After that, the minor version number is increased by 1 in sequence. For example, the second version number is 0.2.0. A version number consists of a major version number, a minor version number, and a stage version number, separated by periods (.). The major version number and minor version number are numeric. The stage version number begins with a digit and can include numbers, underscores, and letters. For example, the version number is 1.1.0 or 2.3.4_beta.
Default: Null
Required: false
Uri:
Type: String
Description:
en: |-
Model version URI, that is, the location where the model is stored. Possible values are:
- The HTTP(S) address of the model, for example: 'https://myweb.com/mymodel.tar.gz '.
- If the model is stored in OSS, the format is 'oss://<bucket>.<endpoint>/object '. For the endpoint configuration, see [Access domain name and data center](~~ 31837 ~~). For example, see OSS:// mybucket/'.
Required: true
TrainingSpec:
Description:
en: Training configuration. Configuration for fine-tuning, incremental training.
Required: false
Default: Null
Type: Json
Options:
Type: String
Description:
en: The extended field. This is a JSON string.
Default: Null
Required: false
ModelId:
Type: String
Description:
en: Model ID. For details about how to obtain the model ID.
Required: true
Resources:
ExtensionResource:
Type: ALIYUN::PAI::ModelVersion
Properties:
VersionDescription:
Ref: VersionDescription
FormatType:
Ref: FormatType
SourceType:
Ref: SourceType
SourceId:
Ref: SourceId
ApprovalStatus:
Ref: ApprovalStatus
ExtraInfo:
Ref: ExtraInfo
FrameworkType:
Ref: FrameworkType
InferenceSpec:
Ref: InferenceSpec
Metrics:
Ref: Metrics
Labels:
Ref: Labels
VersionName:
Ref: VersionName
Uri:
Ref: Uri
TrainingSpec:
Ref: TrainingSpec
Options:
Ref: Options
ModelId:
Ref: ModelId
Outputs:
VersionDescription:
Value:
Fn::GetAtt:
- ExtensionResource
- VersionDescription
Description: Model version description.
GmtModifiedTime:
Value:
Fn::GetAtt:
- ExtensionResource
- GmtModifiedTime
Description: Finally, update the model UTC time in the format iso8601.
FormatType:
Value:
Fn::GetAtt:
- ExtensionResource
- FormatType
Description: Model format, possible values:.
SourceType:
Value:
Fn::GetAtt:
- ExtensionResource
- SourceType
Description: Model source type, possible values:.
SourceId:
Value:
Fn::GetAtt:
- ExtensionResource
- SourceId
Description: Source ID.
ApprovalStatus:
Value:
Fn::GetAtt:
- ExtensionResource
- ApprovalStatus
Description: Admission status, with values as follows:.
ExtraInfo:
Value:
Fn::GetAtt:
- ExtensionResource
- ExtraInfo
Description: Other information.
UserId:
Value:
Fn::GetAtt:
- ExtensionResource
- UserId
Description: The user ID.
FrameworkType:
Value:
Fn::GetAtt:
- ExtensionResource
- FrameworkType
Description: Model framework, possible values:.
InferenceSpec:
Value:
Fn::GetAtt:
- ExtensionResource
- InferenceSpec
Description: Describes how to apply to downstream inference services, such as processors and containers of EAS.
Metrics:
Value:
Fn::GetAtt:
- ExtensionResource
- Metrics
Description: Model indicators.
Labels:
Value:
Fn::GetAtt:
- ExtensionResource
- Labels
Description: List of model version labels.
VersionName:
Value:
Fn::GetAtt:
- ExtensionResource
- VersionName
Description: Model version.
OwnerId:
Value:
Fn::GetAtt:
- ExtensionResource
- OwnerId
Description: The Alibaba Cloud account ID.
Uri:
Value:
Fn::GetAtt:
- ExtensionResource
- Uri
Description: Model version URI, that is, the location where the model is stored.
TrainingSpec:
Value:
Fn::GetAtt:
- ExtensionResource
- TrainingSpec
Description: Training configuration. Configuration for fine-tuning, incremental training.
Options:
Value:
Fn::GetAtt:
- ExtensionResource
- Options
Description: Extension field. The JsonString type.
GmtCreateTime:
Value:
Fn::GetAtt:
- ExtensionResource
- GmtCreateTime
Description: Create a model UTC time in the format ISO8601.
{
"ROSTemplateFormatVersion": "2015-09-01",
"Parameters": {
"VersionDescription": {
"Type": "String",
"Description": {
"en": "Model version description."
},
"Default": null,
"Required": false
},
"FormatType": {
"Type": "String",
"Description": {
"en": "Model format, possible values:\n- OfflineModel\n- SavedModel\n- Keras H5\n- Frozen Pb\n- Caffe Prototxt\n- TorchScript\n- XGBoost\n- PMML\n- AlinkModel\n- ONNX."
},
"AllowedValues": [
"OfflineModel",
"SavedModel",
"Keras H5",
"Frozen Pb",
"Caffe Prototxt",
"TorchScript",
"XGBoost",
"PMML",
"AlinkModel",
"ONNX"
],
"Default": null,
"Required": false
},
"SourceType": {
"Type": "String",
"Description": {
"en": "Model source type, possible values:\n- Custom: Custom.\n- PAIFlow:PAI workflow.\n- TrainingService:PAI training service."
},
"AllowedValues": [
"Custom",
"PAIFlow",
"TrainingService"
],
"Default": null,
"Required": false
},
"SourceId": {
"Type": "String",
"Description": {
"en": "Source ID.\n* When the source type is Custom, this field is not restricted.\n* When the source is PAIFlow or TrainingService, the format is\n'''\nregion=<region_id>,workspaceId=<workspace_id>,kind=<kind>,id=<id>\n'''\nAmong them:\n- region is the Alibaba Cloud region ID.\n- workspaceId indicates the workspace ID.\n- kind: is a type. Value: PipelineRun(PAIFlow workflow);ServiceJob (training service).\n- id: is a unique identifier."
},
"Default": null,
"Required": false
},
"ApprovalStatus": {
"Type": "String",
"Description": {
"en": "Admission status, with values as follows:\n- Pending: to be determined.\n- Approved: Allow to go online.\n- Rejected: Online is not allowed."
},
"AllowedValues": [
"Pending",
"Approved",
"Rejected"
],
"Default": null,
"Required": false
},
"ExtraInfo": {
"Description": {
"en": "The additional information."
},
"Required": false,
"Default": null,
"Type": "Json"
},
"FrameworkType": {
"Type": "String",
"Description": {
"en": "Model framework, possible values:\n- Pytorch\n- XGBoost\n- Keras\n- Caffe\n- Alink\n- Xflow\n- TensorFlow."
},
"AllowedValues": [
"Pytorch",
"XGBoost",
"Keras",
"Caffe",
"Alink",
"Xflow",
"TensorFlow"
],
"Default": null,
"Required": false
},
"InferenceSpec": {
"Description": {
"en": "Describes how to apply to downstream inference services. For example, describe the processor and container of EAS. Example: { \"processor\": \"tensorflow_gpu_1.12\" }"
},
"Required": false,
"Default": null,
"Type": "Json"
},
"Metrics": {
"Description": {
"en": "The metrics for the model. The length after serialization is limited to 8,192. Example:\n{ \"Results\": [{ \"Dataset\": { \"DatasetId\": \"d-sdkjanksaklerhfd\" }, \"Metrics\": { \"cer\": 0.175 } }, { \"Dataset\": { \"Uri\": \"oss://xxxx/\" }, \"Metrics\": { \"cer\": 0.172 } }] }"
},
"Required": false,
"Default": null,
"Type": "Json"
},
"Labels": {
"Description": {
"en": "List of model version labels."
},
"Required": false,
"Default": null,
"Type": "Json",
"MaxLength": 20,
"AssociationProperty": "List[Parameters]",
"AssociationPropertyMetadata": {
"Parameters": {
"Key": {
"Type": "String",
"Description": {
"en": "label key."
},
"Default": null,
"Required": false
},
"Value": {
"Type": "String",
"Description": {
"en": "label value."
},
"Default": null,
"Required": false
}
}
}
},
"VersionName": {
"Type": "String",
"Description": {
"en": "The model version, which is unique for each model. If you leave this parameter empty, the first version is 0.1.0 by default. After that, the minor version number is increased by 1 in sequence. For example, the second version number is 0.2.0. A version number consists of a major version number, a minor version number, and a stage version number, separated by periods (.). The major version number and minor version number are numeric. The stage version number begins with a digit and can include numbers, underscores, and letters. For example, the version number is 1.1.0 or 2.3.4_beta."
},
"Default": null,
"Required": false
},
"Uri": {
"Type": "String",
"Description": {
"en": "Model version URI, that is, the location where the model is stored. Possible values are:\n- The HTTP(S) address of the model, for example: 'https://myweb.com/mymodel.tar.gz '.\n- If the model is stored in OSS, the format is 'oss://<bucket>.<endpoint>/object '. For the endpoint configuration, see [Access domain name and data center](~~ 31837 ~~). For example, see OSS:// mybucket/'."
},
"Required": true
},
"TrainingSpec": {
"Description": {
"en": "Training configuration. Configuration for fine-tuning, incremental training."
},
"Required": false,
"Default": null,
"Type": "Json"
},
"Options": {
"Type": "String",
"Description": {
"en": "The extended field. This is a JSON string."
},
"Default": null,
"Required": false
},
"ModelId": {
"Type": "String",
"Description": {
"en": "Model ID. For details about how to obtain the model ID."
},
"Required": true
}
},
"Resources": {
"ExtensionResource": {
"Type": "ALIYUN::PAI::ModelVersion",
"Properties": {
"VersionDescription": {
"Ref": "VersionDescription"
},
"FormatType": {
"Ref": "FormatType"
},
"SourceType": {
"Ref": "SourceType"
},
"SourceId": {
"Ref": "SourceId"
},
"ApprovalStatus": {
"Ref": "ApprovalStatus"
},
"ExtraInfo": {
"Ref": "ExtraInfo"
},
"FrameworkType": {
"Ref": "FrameworkType"
},
"InferenceSpec": {
"Ref": "InferenceSpec"
},
"Metrics": {
"Ref": "Metrics"
},
"Labels": {
"Ref": "Labels"
},
"VersionName": {
"Ref": "VersionName"
},
"Uri": {
"Ref": "Uri"
},
"TrainingSpec": {
"Ref": "TrainingSpec"
},
"Options": {
"Ref": "Options"
},
"ModelId": {
"Ref": "ModelId"
}
}
}
},
"Outputs": {
"VersionDescription": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"VersionDescription"
]
},
"Description": "Model version description."
},
"GmtModifiedTime": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"GmtModifiedTime"
]
},
"Description": "Finally, update the model UTC time in the format iso8601."
},
"FormatType": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"FormatType"
]
},
"Description": "Model format, possible values:."
},
"SourceType": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"SourceType"
]
},
"Description": "Model source type, possible values:."
},
"SourceId": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"SourceId"
]
},
"Description": "Source ID."
},
"ApprovalStatus": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"ApprovalStatus"
]
},
"Description": "Admission status, with values as follows:."
},
"ExtraInfo": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"ExtraInfo"
]
},
"Description": "Other information."
},
"UserId": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"UserId"
]
},
"Description": "The user ID."
},
"FrameworkType": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"FrameworkType"
]
},
"Description": "Model framework, possible values:."
},
"InferenceSpec": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"InferenceSpec"
]
},
"Description": "Describes how to apply to downstream inference services, such as processors and containers of EAS."
},
"Metrics": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"Metrics"
]
},
"Description": "Model indicators."
},
"Labels": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"Labels"
]
},
"Description": "List of model version labels."
},
"VersionName": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"VersionName"
]
},
"Description": "Model version."
},
"OwnerId": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"OwnerId"
]
},
"Description": "The Alibaba Cloud account ID."
},
"Uri": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"Uri"
]
},
"Description": "Model version URI, that is, the location where the model is stored."
},
"TrainingSpec": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"TrainingSpec"
]
},
"Description": "Training configuration. Configuration for fine-tuning, incremental training."
},
"Options": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"Options"
]
},
"Description": "Extension field. The JsonString type."
},
"GmtCreateTime": {
"Value": {
"Fn::GetAtt": [
"ExtensionResource",
"GmtCreateTime"
]
},
"Description": "Create a model UTC time in the format ISO8601."
}
}
}该文章对您有帮助吗？