What are the different stages of dataset training and where can I find them?

If you have triggered the training of an AI Model, you are able to get updates on the status of the training using the right menu bar in the "AI Models" menu. In the AI Models view, you see the status of all current and past trainings on the right-hand-side.

What information does it provide?

All training jobs are clustered in three categories:

• Processing, Completed or Failed.

Each Job card features the following information:

• Training ID Number
• Model Name (Result of the Training)
• Dataset Name (Source of the Training)
• Start Time of the Training
• Completion Time of the Training
• Duration of the Training 

In case the Training has failed, there are two more options:

• Failure Time
• Failure Reason (with short description for troubleshooting)

What are the possible training statuses for a dataset?

• Good:
  The dataset contains no errors or warnings and is ready for AI model training.

• Pending:
  The training of a new AI model has been initiated but has not started processing yet.

• Processing:
  The AI model is currently being trained and is in progress.

• Completed:
  The training process is finished, and the AI model has been successfully created. It can now be used in analysis.

• Failed:
  The training process was unsuccessful, and no AI model was created. This could be due to:

  • Insufficient images (minimum of three for Face Recognition or one for Landmark Recognition).
  • The specified dataset does not exist (API issue).
  • Technical problems with DeepVA servers.