Clarification on map metric

What map does TLT use by default? Just AP50, or mAp from 0.5 to 0.75? Or mAp from 0.5 to 0.95?

Which network are you talking about?

Any detection network in transfer learning toolkit. For example, yolov3.

I mean when evaluating the network

See https://docs.nvidia.com/metropolis/TLT/tlt-getting-started-guide/text/evaluating_model.html

The classification app computes evaluation loss, Top-k accuracy, precision and recall as metrics. Meanwhile, tlt-evaluate for DetectNet_v2, FasterRCNN, Retinanet, DSSD, YOLOV3, and SSD computes the Average Precision per class and the mean Average Precision metrics as defined in the Pascal VOC challenge. Both sample and integrate modes are supported to calculate average precision. The former was used in VOC challenges before 2010 while the latter was used from 2010 onwards. The SAMPLE mode uses an 11-point method to compute the AP, while the INTEGRATE mode uses a more fine-grained integration method and gets a more accurate number of AP. MaskRCNN reports COCO’s detection evaluation metrics. AP50 in COCO metrics is comparable to mAP in Pascal VOC metrics.