View a prediction's tabs

Overview

A prediction contains the following tabs that you can view when the prediction is running or completed:

• Summary
• Metrics
• Test Prediction Insights
• Test Interpretation Insights
• Logs
• Config

Instructions

To access the tabs of a completed or running prediction, consider the following instructions:

1. Open the predictions list

   In the H2O Hydrogen Torch navigation menu, click View predictions.

2. Open the prediction

   Click the name of the prediction.

Tabs

Summary

On the Summary tab, H2O Hydrogen Torch displays the following information about the selected prediction:

Name	Description
Name	Name of the prediction.
Dataset	Dataset used for the prediction.
Problem type	Problem type of the experiment.
Fold	The validation fold value.
Seed	The random seed value that you selected for model training.
Gpu list	List of GPUs that H2O Hydrogen Torch used to run the prediction.
Loss	The loss function that the original experiment optimized during model training.
Metric	The metric selected to evaluate model performance.
Train metric	The evaluation metric computed on the training data. Predictions do not run a training pass, so this value is empty.
Val metric	The evaluation metric computed on the validation data. Predictions do not run a validation pass, so this value is empty.
Test metric	The evaluation metric computed on the test data. Available only when the test dataset contains labels.
Batch inference speed	The average wall-clock time per batch while scoring the test data, including prediction postprocessing.

Metrics

On the Metrics tab, H2O Hydrogen Torch displays all available test metrics applicable to the problem type of the prediction, if the test dataset contains labels.

note

Precision, recall, F1, F05, and F2 are micro-average. To learn more, see Metrics.

Test prediction insights

On the Test prediction insights tab, H2O Hydrogen Torch visualizes test samples and their prediction. In addition, if the test dataset contains labels, visualizations are divided as follows:

• Random test samples
  • In the Random test samples tab, H2O Hydrogen Torch visualizes best and worst test samples and their prediction.
  Note

  Each random test sample contains a Loss value.

• Best test samples
  • In the Best test samples tab, H2O Hydrogen Torch visualizes the best test samples and their prediction. Samples in this tab can also be regarded as the model's smallest significant errors.
  Note

  Each best test sample contains a Loss value.

• Worst test samples
  • In the Worst test samples tab, H2O Hydrogen Torch visualizes the worst test samples and their prediction. Samples in this tab can also be regarded as the model's most significant errors. Samples in this tab can help identify issues the model might have.
  Note

  Each worst test sample contains a Loss value.

Note

Evaluating the best and the worst performance differs for each problem type and might vary from the selected loss and the selected metric to assess the model.

Test interpretation insights

On the Test Interpretation Insights tab, H2O Hydrogen Torch displays interpretation insights for the test samples of the prediction. The tab appears only when the prediction has interpretation insights; H2O Hydrogen Torch supports them for image, text, and audio classification and regression problem types. To learn more about interpretation insights, see Validation interpretation insights.

Logs

On the Logs tab, you can view the logs of the prediction.

Config

On the Config tab, you can view the experiment configurations used to generate the prediction.

---

Feedback

• Submit and view feedback for this page
• Send feedback about H2O Hydrogen Torch to cloud-feedback@h2o.ai