Help Center/ ModelArts/ SDK Reference/ Training Management (Old Version)/ Training Job Versions/ Querying the Details About a Training Job Version
Updated on 2024-03-21 GMT+08:00

Querying the Details About a Training Job Version

Sample Code

In ModelArts notebook, you do not need to enter authentication parameters for session authentication. For details about session authentication of other development environments, see Session Authentication.

  • Method 1: Use the specified job_id.
    1
    2
    3
    4
    5
    6
    from modelarts.session import Session
    from modelarts.estimator import Estimator
    session = Session()
    estimator = Estimator(session, job_id="182626")
    job_version_info = estimator.get_job_version_info()
    print(job_version_info)
    
  • Method 2: Use the training job version created in Creating a Training Job Version.
    1
    2
    job_version_info = job_version_instance.get_job_version_info()
    print(job_version_info)
    

Parameters

Table 1 Estimator request parameters

Parameter

Mandatory

Type

Description

modelarts_session

Yes

Object

Session object. For details about the initialization method, see Session Authentication.

job_id

Yes

String

ID of a training job. You can query job_id using the training job object generated in Creating a Training Job, for example, job_instance.job_id, or from the response obtained in Querying the List of Training Jobs.

Table 2 get_job_version_info response parameters

Parameter

Type

Description

error_msg

String

Error message when the API call fails.

This parameter is not included when the API call succeeds.

error_code

String

Error code when the API fails to be called. For details, see Error Codes.

This parameter is not included when the API call succeeds.

is_success

Boolean

Whether the API call succeeds

job_id

Long

Training job ID

job_name

String

Training job name

job_desc

String

Description of a training job

version_count

Long

Number of versions of a training job

versions

JSON Array

Version parameters of a training job

Table 3 versions parameters

Parameter

Type

Description

version_id

Long

Version ID of a training job

version_name

String

Version name of a training job

pre_version_id

Long

ID of the previous version of a training job

engine_type

Long

Engine type of a training job

engine_id

Long

ID of the engine selected for a training job

engine_version

String

Version of the engine selected for a training job

status

Integer

Status of a training job

app_url

String

Code directory of a training job

boot_file_url

String

Boot file of a training job

create_time

Long

Time when a training job is created

parameter

JSON Array

Running parameters of a training job. It is a collection of label-value pairs. This parameter is a container environment variable when a job uses a custom image.

duration

Long

Training job running duration, in milliseconds

spec_id

Long

ID of the resource specifications selected for a training job

core

String

Number of cores of the resource specifications

cpu

String

CPU memory of the resource specifications

gpu_num

Integer

Number of GPUs of the resource specifications

gpu_type

String

GPU type of the resource specifications

worker_server_num

Integer

Number of workers in a training job

data_url

String

Dataset of a training job

train_url

String

OBS path to the training job output file

log_url

String

OBS URL of the logs of a training job. By default, this parameter is left blank. Example value: /usr/log/

dataset_version_id

String

Dataset version ID of a training job

dataset_id

String

Dataset ID of a training job

data_source

JSON Array

Datasets of a training job

model_id

String

Model ID of a training job

model_metric_list

JSON Array

Model metrics of a training job

system_metric_list

JSON Array

System monitoring metrics of a training job

user_image_url

String

SWR URL of the custom image used by a training job

user_command

String

Boot command used to start the container of the custom image of a training job

Table 4 data_source parameters

Parameter

Type

Description

dataset_id

String

Dataset ID of a training job

dataset_version

String

Dataset version ID of a training job

type

String

Dataset type

obs: Data from OBS is used.

dataset: Data from a specified dataset is used.

data_url

String

OBS bucket path

Table 5 model_metric_list parameters

Parameter

Type

Description

metric

JSON Array

Validation metrics of a class of a training job

total_metric

JSON Array

All validation metrics of a training job

Table 6 system_metric_list parameters

Parameter

Type

Description

cpuUsage

JSON Array

CPU usage of a training job

memUsage

JSON Array

Memory usage of a training job

gpuUtil

JSON Array

GPU usage of a training job

Table 7 metric parameters

Parameter

Type

Description

metric_values

JSON Array

Validation metrics of a class of a training job

reserved_data

JSON Array

Reserved parameter

metric_meta

JSON Array

A class of a training job, including the class ID and name

Table 8 metric_values parameters

Parameter

Type

Description

recall

JSON Array

Recall of a class of a training job

precision

JSON Array

Precision of a class of a training job

accuracy

JSON Array

Accuracy of a class of a training job

Table 9 total_metric parameters

Parameter

Type

Description

total_metric_meta

JSON Array

Reserved parameter

total_reserved_data

JSON Array

Reserved parameter

total_metric_values

JSON Array

All validation metrics of a training job

Table 10 total_metric_values parameters

Parameter

Type

Description

f1_score

Float

F1 score of a training job

recall

Float

Total recall of a training job

precision

Float

Total precision of a training job

accuracy

Float

Total accuracy of a training job