Fix incorrect/unclear sensitivity parameter help string

Author: MatthewScholefield

precise/scripts/train.py

@@ -29,7 +29,8 @@
     Number of epochs to train model for
 
 :-s --sensitivity float 0.2
-    Weighted loss bias. Higher values decrease increase positives
+    Target sensitivity when training. Higher values cause
+    more false positives
 
 :-b --batch-size int 5000
     Batch size for training

precise/scripts/train_generated.py

@@ -35,7 +35,8 @@
     random audio that should not cause an activation
 
 :-s --sensitivity float 0.2
-    Weighted loss bias. Higher values decrease increase positives
+    Target sensitivity when training. Higher values cause
+    more false positives
 
 :-sb --save-best
     Only save the model each epoch if its stats improve