Here you can test the speech model.
Sam will test the model set by in the Output files section in the input & output section.
During testing, the recognizer of the selected backend will run to recognize on the input files as set by the test prompts. Resulting word- and sentence errors will be counted as well as the overall robustness by analyzing confidence scores of the recognizer.
The main test screen will show all started tests and report when they complete. The tests themselves will run threaded (at the same time). The amount of parallelism best suited for your machine is determined automatically.
As soon as all tests are finished, the test results will be displayed automatically.