Fix for AP calculation limits 0.0 - 1.0 (#4563)

This PR brings alignment in AP computation practices with Detectron2 and MMDetection. Problem first noted by @yusiyoh in https://github.com/ultralytics/yolov5/issues/4546

Fix for AP calculation limits 0.0 - 1.0 (#4563)
8ac96b79 · Glenn Jocher · GitHub · e899d6e8 · 8ac96b79
--- a/utils/metrics.py
+++ b/utils/metrics.py
@@ -91,8 +91,8 @@ def compute_ap(recall, precision):
    """

    # Append sentinel values to beginning and end
-    mrec = np.concatenate(([0.], recall, [recall[-1] + 0.01]))
-    mpre = np.concatenate(([1.], precision, [0.]))
+    mrec = np.concatenate(([0.0], recall, [1.0]))
+    mpre = np.concatenate(([1.0], precision, [0.0]))

    # Compute the precision envelope
    mpre = np.flip(np.maximum.accumulate(np.flip(mpre)))