Syntiant Tiny ML inferencing not working correctly

Help
#1

Question/Issue:
When I create a library to deploy a keyword spotting model locally on the Syntiant board, the posterior parameters that I modified in the Deployment tab do not seem to change. For example, when I change the parameter ‘phth’ to 0.01 for a keyword, it does not seem to trigger a prediction. I would expect it to trigger almost constantly while I’m not saying anything as the threshold is so low.

Project ID:
157724

#2

Also, what does ‘Performance Calibration’ do for me? When I adjust the config to get a FAR of 0% and save it, where can I see that it adjusted the model that I’m going to deploy? It does not seem to be present in the posterior parameters as well. I do not get the idea that the performance calibration is included in the model. Is this the case?