创建任务

Path Parameters

subscription_name string required

resource_group_name string required

资源组

zone string required

可用区. 工作空间可用区，一般为region+z，例如：上海的region为cn-sh-01，工作空间的可用区则为cn-sh-01z

workspace_name string required

工作空间名称

Query Parameters

training_job_name string

训练任务名称

application/json

Request Body required

name string

可选,任务名称. 不填写，则由后台默认随机生成. 若填入,则需要遵循以下规范: 1、1～63个字符 2、小写字母开头，小写字母或数字结尾 3、可包含小写英文字母、数字、中划线-

display_name string required

任务显示名. 需遵循规范: 1、1-63个字符 2、以中文字符、英文字母或数字开头 3、可包含中文字符、英文字母、数字、下划线“_"和连字符"-"

ownership object

所有权关系. 只输出,请求无需传入

framework enum required

Possible values: [FRAMEWORK_UNSPECIFIED, MPI, PYTORCH, PYTORCH_DDP, TENSORFLOW, TENSORFLOW_PS, SENSEPARROTS, SENSEPARROTS_DDP]

框架类型

roles object[] required

角色. 当前仅支持填写单个role

Array [

name string required

角色名. 当前只支持填写Worker

resource_spec object[] required

资源规格

Array [

name string required

资源规格名称. 根据资源规格列表获取. 例: N2lS.Ie.I60.2

replicas int32

请求调度副本数. 当前规格所需 worker 的数量. 当前不支持手动强异构调度，暂不支持

limits object

资源列表. key值: CPU(cpu)、内存(memory)、设备数量(device)

property name* string

requests object

资源列表. key值: CPU(cpu)、内存(memory)、设备数量(device)

property name* string

]

total_replicas int32

总副本. 所需 worker 的数量. 选择自动弱异构调度或者单规格时，total_replicas 必传

startup_script string required

启动脚本

image_path string required

镜像真实uri path

]

env object[]

可选,环境变量配置,最多支持10个

Array [

key string required

key

value string required

value

]

mount object[]

卷挂载

Array [

type enum required

Possible values: [MOUNT_TYPE_UNSPECIFIED, PV_AFS, PV_ACS]

挂载类型

id string required

存储 id

mount_path string required

挂载路径

subdir string

指定挂载A卷下的子目录

]

tensorboard object

TensorBoard配置

log_path string

开启tensorboard之后的日志路径，需要与任一存储卷挂载路径保持一致. 是否开启tensorboard，根据路径值来判断

scheduling object

调度相关

priority enum

Possible values: [PRIORITY_UNSPECIFIED, NORMAL, HIGH, HIGHEST]

调度优先级, 默认为: NORMAL

quota_type enum

Possible values: [JOB_QUOTA_TYPE_UNSPECIFIED, RESERVED, SPOT]

可选, 任务闲时配额. 默认为: RESERVED 当前支持RESERVED和SPOT两种只有集群开启显示配额,才能创建闲时资源(spot)任务只有集群和工作空间的关联关系为全部时,才能创建标准(reserved)任务

resource_pool object required

资源池

name string required

资源名称. 例: 集群名称、预留配额名称等,需要和资源类型对应上

zone string

公共集群的可用区,仅当使用公共集群创建任务时,需要传入,内容要与vpc_id对应上

vpc_id string

vpc id. 使用公共集群创建任务时,需要指定vpc的id

fine_tuning_specs string

[beta]模型微调配置项[json string]

async_checkpoint object

[beta]AsyncCheckpoint. 是否开启，根据最大缓存轮数来判断

max_ckpt_rounds int32

[beta]最大缓存轮数

lme object

[beta]大模型可配置项

enable_warmingup boolean

[beta]任务运行前是否开启环境诊断

max_retries int32

[beta]配置重试最大次数

enable_health_monitor boolean

[beta]是否开启健康检测

Responses

200
default

application/json

Schema

Example (from schema)

Schema

name string

display_name string required

任务显示名. 需遵循规范: 1、1-63个字符 2、以中文字符、英文字母或数字开头 3、可包含中文字符、英文字母、数字、下划线“_"和连字符"-"

ownership object

所有权关系. 只输出,请求无需传入

framework enum required

Possible values: [FRAMEWORK_UNSPECIFIED, MPI, PYTORCH, PYTORCH_DDP, TENSORFLOW, TENSORFLOW_PS, SENSEPARROTS, SENSEPARROTS_DDP]

框架类型

roles object[] required

角色. 当前仅支持填写单个role

Array [

name string required

角色名. 当前只支持填写Worker

resource_spec object[] required

资源规格

Array [

name string required

资源规格名称. 根据资源规格列表获取. 例: N2lS.Ie.I60.2

replicas int32

请求调度副本数. 当前规格所需 worker 的数量. 当前不支持手动强异构调度，暂不支持

limits object

资源列表. key值: CPU(cpu)、内存(memory)、设备数量(device)

property name* string

requests object

资源列表. key值: CPU(cpu)、内存(memory)、设备数量(device)

property name* string

]

total_replicas int32

总副本. 所需 worker 的数量. 选择自动弱异构调度或者单规格时，total_replicas 必传

startup_script string required

启动脚本

image_path string required

镜像真实uri path

]

env object[]

可选,环境变量配置,最多支持10个

Array [

key string required

key

value string required

value

]

mount object[]

卷挂载

Array [

type enum required

Possible values: [MOUNT_TYPE_UNSPECIFIED, PV_AFS, PV_ACS]

挂载类型

id string required

存储 id

mount_path string required

挂载路径

subdir string

指定挂载A卷下的子目录

]

tensorboard object

TensorBoard配置

log_path string

开启tensorboard之后的日志路径，需要与任一存储卷挂载路径保持一致. 是否开启tensorboard，根据路径值来判断

scheduling object

调度相关

priority enum

Possible values: [PRIORITY_UNSPECIFIED, NORMAL, HIGH, HIGHEST]

调度优先级, 默认为: NORMAL

quota_type enum

Possible values: [JOB_QUOTA_TYPE_UNSPECIFIED, RESERVED, SPOT]

resource_pool object required

资源池

name string required

资源名称. 例: 集群名称、预留配额名称等,需要和资源类型对应上

zone string

公共集群的可用区,仅当使用公共集群创建任务时,需要传入,内容要与vpc_id对应上

vpc_id string

vpc id. 使用公共集群创建任务时,需要指定vpc的id

fine_tuning_specs string

[beta]模型微调配置项[json string]

async_checkpoint object

[beta]AsyncCheckpoint. 是否开启，根据最大缓存轮数来判断

max_ckpt_rounds int32

[beta]最大缓存轮数

lme object

[beta]大模型可配置项

enable_warmingup boolean

[beta]任务运行前是否开启环境诊断

max_retries int32

[beta]配置重试最大次数

enable_health_monitor boolean

[beta]是否开启健康检测

{
  "subscription_name": "string",
  "resource_group_name": "string",
  "zone": "string",
  "workspace_name": "string",
  "name": "string",
  "uid": "string",
  "display_name": "string",
  "id": "string",
  "ownership": {
    "user_name": "string",
    "user_id": "string",
    "requestor_id": "string",
    "tenant_id": "string"
  },
  "framework": "FRAMEWORK_UNSPECIFIED",
  "roles": [
    {
      "name": "string",
      "resource_spec": [
        {
          "name": "string",
          "replicas": 0,
          "description": "string",
          "limits": {},
          "requests": {}
        }
      ],
      "total_replicas": 0,
      "startup_script": "string",
      "image_path": "string"
    }
  ],
  "env": [
    {
      "key": "string",
      "value": "string"
    }
  ],
  "mount": [
    {
      "type": "MOUNT_TYPE_UNSPECIFIED",
      "id": "string",
      "mount_path": "string",
      "subdir": "string"
    }
  ],
  "tensorboard": {
    "log_path": "string"
  },
  "scheduling": {
    "priority": "PRIORITY_UNSPECIFIED",
    "quota_type": "JOB_QUOTA_TYPE_UNSPECIFIED"
  },
  "resource_pool": {
    "name": "string",
    "display_name": "string",
    "zone": "string",
    "vpc_id": "string"
  },
  "state": "STATE_UNSPECIFIED",
  "root_mapping": true,
  "create_time": "2024-11-04T06:28:49.172Z",
  "start_time": "2024-11-04T06:28:49.172Z",
  "suspend_time": "2024-11-04T06:28:49.172Z",
  "update_time": "2024-11-04T06:28:49.172Z",
  "complete_time": "2024-11-04T06:28:49.172Z",
  "delete_time": "2024-11-04T06:28:49.172Z",
  "fine_tuning_specs": "string",
  "async_checkpoint": {
    "max_ckpt_rounds": 0
  },
  "lme": {
    "enable_warmingup": true,
    "max_retries": 0,
    "enable_health_monitor": true,
    "current_retries": 0
  }
}

Default error response

application/json

Schema

Example (from schema)

Schema

code int32

The status code, which should be an enum value of [google.rpc.Code][google.rpc.Code].

message string

A developer-facing error message, which should be in English. Any user-facing error message should be localized and sent in the [google.rpc.Status.details][google.rpc.Status.details] field, or localized by the client.

details object[]

A list of messages that carry the error details. There is a common set of message types for APIs to use.

Array [

@type string

The type of the serialized message.

]

{
  "code": 0,
  "message": "string",
  "details": [
    {
      "@type": "string"
    }
  ]
}

创建任务​

创建任务