Skip to main content

Score threshold's effect on different eval metrics

Below are plotted precision, recall and mAP for both train and test data for different score threshold values (calibration parameter of the inference model). Score threshold values above 0.1 seem to give better trade off between precision and recall for both data sets.
Created on December 29|Last edited on December 29

Section 1

Add markdown, images, and LaTeX\LaTeX




0.10.20.30.40.5score_threshold0.760.780.80.820.840.860.88test_recall_person
0.10.20.30.40.5score_threshold0.20.40.60.8test_precision_person
0.10.20.30.40.5score_threshold0.760.780.80.820.840.86test_mAP_person
Sweep: lw526gt7 1
20
Sweep: lw526gt7 2
0