Updated on 2024-05-07 GMT+08:00

Viewing Training Job Details

  1. Log in to the ModelArts management console.
  2. In the navigation pane on the left, choose Training Management > Training Jobs.
  3. In the training job list, click a job name to switch to the training job details page.
  4. On the left of the training job details page, view basic job settings and algorithm parameters.
    • Basic job settings
      Table 1 Basic job settings

      Parameter

      Description

      Job ID

      Unique ID of a training job

      Status

      Training job status

      Created

      Time when the training job is created

      Duration

      Running duration of a training job

      Retries

      Number of times that a training job automatically restarts upon a fault. This parameter is available only when Auto Restart is enabled during training job creation.

      Description

      Description of a training job.

      You can click the edit icon to update the description of a training job.

    • Algorithm parameters
      Table 2 Algorithm parameters

      Parameter

      Description

      Algorithm Name

      Algorithm used in a training job You can click the algorithm name to go to the algorithm details page.

      Preset images

      Preset image used by a training job

      Code Directory

      OBS path to the code directory of a training job

      You can click Edit Code on the right to edit the training script code in OBS Online Editor. OBS Online Editor is not available for a training job in the Pending, Creating, or Running status.

      NOTE:

      If you use the algorithm subscribed in AI Gallery to create a training job, then this parameter is not supported.

      Boot File

      Location where a boot file is stored.

      NOTE:

      If you use the algorithm subscribed in AI Gallery to create a training job, then this parameter is not supported.

      User ID

      ID of the user who runs the container.

      Local Code Directory

      Path to the training code in the training container

      Work Directory

      Path to the training startup file in the training container

      Compute Nodes

      Number of compute nodes

      Dedicated resource pool

      Dedicated resource pool information. This parameter is available only when a training job uses a dedicated resource pool.

      Specifications

      Training specifications used in a training job

      Input - Input Path

      OBS path where the input data is stored

      Input - Parameter Name

      Algorithm code parameter specified by the input path

      Input - Obtained from

      Method of obtaining the training job input.

      Input - Local Path (Training Parameter Value)

      Path for storing the input data in the ModelArts backend container. After the training is started, ModelArts downloads the data stored in OBS to the backend container.

      Output - Output Path

      OBS path where the output data is stored

      Output - Parameter Name

      Algorithm code parameter specified by the output path

      Output - Obtained from

      Method of obtaining the training job output.

      Output - Local Path (Training Parameter Value)

      Path for storing the output data in the ModelArts backend container

      Hyperparameter

      Hyperparameters used in a training job

      Environment Variable

      Environment variables for a training job