Viewing Job Details

After a training job finishes, you can manage the training job versions and check whether the training result of the job is satisfactory by viewing the job details and Viewing the Evaluation Result.

Training Job Details

In the left navigation pane of the ModelArts management console, choose Training Management > Training Jobs to switch to the Training Jobs page. In the training job list, click a job name to view the job details.

Table 1 lists parameters of the training job of each version.

Figure 1 Training job details

**Table 1** Training job details
Parameter	Description
Version	Version of a training job, which is automatically defined by the system, for example, V0001 and V0002.
Status	Status of a training job,
Duration	Running duration of a training job
Configurations	Details about the parameters of the current training job version
Logs	Logs of the current training job version. If you set Log Output Path when creating a training job, you can click the download button on the Logs tab page to download the logs stored in the OBS bucket to the local host.
Resource Usages	Usage of resources of the current training version, including CPUs, GPUs, NPUs, and memory

Viewing the Evaluation Result

When you use a built-in algorithm published by ModelArts to create a training job, you can view the evaluation result of the training job. If the evaluation code is added to the training script based on the ModelArts specifications, you can view the evaluation result on the job details page after the training job is complete. For details about how to add the evaluation code, see Adding the Evaluation Code.

You can view evaluation results of models using algorithms such as Image Classification-ResNet_v1_50 , Object Detection-FasterRCNN_ResNet50, and Object Detection-EfficientDet built-in algorithms.

After a training job is executed and its status is Successful, you can view the evaluation details on the Evaluation Result tab page of the job details page. See Figure 2.

The evaluation results include the evaluation overview, precision evaluation, sensitivity analysis, and common model indicators. The system automatically provides optimization suggestions based on your model metrics. Read the suggestions and guidance on the page carefully to further optimize your model.

Figure 2 Viewing the evaluation result