Description
The max_test_steps
parameter specifies the maximum number of test steps to run after all experiments have been completed. These test steps evaluate the final selected model on the test set to compute metrics that indicate the model's overall performance.
Increasing max_test_steps
enables Kumo to evaluate the model on a larger portion of the test set. This provides a more accurate and reliable assessment of the model's overall performance. However, this comes at the cost of increased runtime.
Reducing max_test_steps
reduces testing time by limiting the number of test samples evaluated. While this speeds up the evaluation process on the test set, it may lead to less representative performance metrics.
For detailed pseudo code illustrating the training and evaluation process, see the max_epochs
documentation.
Supported Task Types
- All
Default Values
run_mode | Default Value |
---|---|
FAST | 2000 |
NORMAL | 4000 |
BEST | 4000 |
Updated 23 days ago