Skip to content

Southern Sierra Nevada Model

The Southern Sierra Nevada model has been created using almost 17 hours of soundscape data containing 10,296 bounding box labels representing 21 different species.


Training Results

The following figure illustrates various data that has been recorded during training. The metrics have been computed after each epoch with the evaluation split of the dataset.

Training Results

Model Checkpoint

The weights that produced the highest metrics/mAP50-95 on the validation split were selected for the final model.


Species Distribution Across Splits

The following table shows the amount of annotations in total and for each species as described in Dataset Splits.

Species Train Val Test Total 70/15/15 Quality
Total 16,528 3,600 3,400 23,528 70.2/15.3/14.5
???? 1,003 214 184 1,401 71.6/15.3/13.1
amepip 2,757 581 637 3,975 69.4/14.6/16.0
amerob 215 45 45 305 70.5/14.8/14.8
clanut 709 151 149 1,009 70.3/15.0/14.8
daejun 368 70 81 519 70.9/13.5/15.6
gcrfin 1,889 411 372 2,672 70.7/15.4/13.9
herthr 508 108 104 720 70.6/15.0/14.4
mouchi 111 23 23 157 70.7/14.6/14.6
rocwre 1,498 319 326 2,143 69.9/14.9/15.2
whcspa 7,119 1,678 1,479 10,276 69.3/16.3/14.4
brebla 2 0 0 2 Train-only
casfin 59 0 0 59 Train-only
dusfly 7 0 0 7 Train-only
foxspa 64 0 0 64 Train-only
mallar3 6 0 0 6 Train-only
moublu 65 0 0 65 Train-only
norfli 21 0 0 21 Train-only
orcwar 16 0 0 16 Train-only
sposan 53 0 0 53 Train-only
warvir 18 0 0 18 Train-only
yelwar 1 0 0 1 Train-only
yerwar 39 0 0 39 Train-only