Skip to content

Southwestern Amazon Basin Model

The Southwestern Amazon Basin model has been created using 21 hours of soundscape data containing 14,798 bounding box labels representing 132 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 79,086 17,126 16,889 113,101 69.9/15.1/14.9
???? 5,481 1,193 1,152 7,826 70.0/15.2/14.7
amabaw1 391 85 89 565 69.2/15.0/15.8
amapyo1 303 68 63 434 69.8/15.7/14.5
baffal1 868 195 190 1,253 69.3/15.6/15.2
bartin2 238 57 50 345 69.0/16.5/14.5
blbthr1 983 209 224 1,416 69.4/14.8/15.8
blcbec1 113 24 28 165 68.5/14.5/17.0
blfant1 4,185 902 906 5,993 69.8/15.1/15.1
blfcot1 98 22 22 142 69.0/15.5/15.5
blfjac1 212 76 54 342 62.0/22.2/15.8
blfnun1 187 41 41 269 69.5/15.2/15.2
bltant2 177 24 29 230 77.0/10.4/12.6
blttro1 634 133 141 908 69.8/14.6/15.5
bsbeye1 198 49 43 290 68.3/16.9/14.8
bubgro2 1,066 230 221 1,517 70.3/15.2/14.6
bubwre1 3,045 654 669 4,368 69.7/15.0/15.3
bucmot4 1,498 319 328 2,145 69.8/14.9/15.3
buffal1 148 27 41 216 68.5/12.5/19.0
butsal1 579 122 136 837 69.2/14.6/16.2
butwoo1 1,744 381 379 2,504 69.6/15.2/15.1
chwfog1 139 32 39 210 66.2/15.2/18.6
cintin1 1,458 304 306 2,068 70.5/14.7/14.8
citwoo1 649 133 145 927 70.0/14.3/15.6
coffal1 111 16 26 153 72.5/10.5/17.0
coltro1 479 95 110 684 70.0/13.9/16.1
cowpar1 231 64 61 356 64.9/18.0/17.1
crfgle1 149 29 29 207 72.0/14.0/14.0
ducatt1 92 32 16 140 65.7/22.9/11.4
ducfly 174 44 33 251 69.3/17.5/13.1
ducgre1 1,310 308 364 1,982 66.1/15.5/18.4
dutant2 728 182 161 1,071 68.0/17.0/15.0
elewoo1 119 29 32 180 66.1/16.1/17.8
fasant1 563 149 124 836 67.3/17.8/14.8
fepowl 121 25 29 175 69.1/14.3/16.6
gilbar1 63 31 14 108 58.3/28.7/13.0
gocspa1 272 61 44 377 72.1/16.2/11.7
goeant1 778 162 151 1,091 71.3/14.8/13.8
grasal3 346 72 72 490 70.6/14.7/14.7
greant1 370 81 86 537 68.9/15.1/16.0
greibi1 97 39 12 148 65.5/26.4/8.1
grfdov1 2,478 535 579 3,592 69.0/14.9/16.1
gryant1 179 36 43 258 69.4/14.0/16.7
gryant2 1,426 327 331 2,084 68.4/15.7/15.9
hauthr1 12,374 2,843 2,845 18,062 68.5/15.7/15.8
horscr1 259 52 54 365 71.0/14.2/14.8
littin1 2,077 436 437 2,950 70.4/14.8/14.8
lobwoo1 518 117 109 744 69.6/15.7/14.7
lowant1 461 92 92 645 71.5/14.3/14.3
meapar 1,430 316 287 2,033 70.3/15.5/14.1
partan1 453 105 102 660 68.6/15.9/15.5
pirfly1 3,726 870 794 5,390 69.1/16.1/14.7
pluant1 1,082 226 234 1,542 70.2/14.7/15.2
plupig2 1,394 318 301 2,013 69.2/15.8/15.0
plwant1 464 104 100 668 69.5/15.6/15.0
putfru1 498 107 102 707 70.4/15.1/14.4
pygant1 298 63 66 427 69.8/14.8/15.5
rebmac2 110 28 20 158 69.6/17.7/12.7
rinant2 122 27 26 175 69.7/15.4/14.9
rinwoo1 83 17 19 119 69.7/14.3/16.0
ruboro1 1,050 219 216 1,485 70.7/14.7/14.5
rudpig 539 135 112 786 68.6/17.2/14.2
rufant3 842 178 176 1,196 70.4/14.9/14.7
ruqdov 181 32 43 256 70.7/12.5/16.8
scrpih1 2,215 469 458 3,142 70.5/14.9/14.6
sobcac1 361 92 80 533 67.7/17.3/15.0
spigua1 167 41 36 244 68.4/16.8/14.8
spwant2 1,262 281 277 1,820 69.3/15.4/15.2
strcuc1 278 36 0 314 88.5/11.5/0.0
strwoo2 537 125 116 778 69.0/16.1/14.9
tabsco1 165 36 37 238 69.3/15.1/15.5
thlwre1 3,958 850 872 5,680 69.7/15.0/15.4
undtin1 1,225 263 261 1,749 70.0/15.0/14.9
viotro3 465 90 88 643 72.3/14.0/13.7
wespuf1 72 15 16 103 69.9/14.6/15.5
whbtot1 118 23 29 170 69.4/13.5/17.1
whnrob1 512 117 117 746 68.6/15.7/15.7
whrsir1 120 27 26 173 69.4/15.6/15.0
whttou1 1,727 374 358 2,459 70.2/15.2/14.6
whwbec1 1,258 274 243 1,775 70.9/15.4/13.7
wibpip1 293 68 67 428 68.5/15.9/15.7
yemfly1 219 48 51 318 68.9/15.1/16.0
yercac1 96 15 9 120 80.0/12.5/7.5
astgna1 42 0 0 42 Train-only
barant1 4 0 0 4 Train-only
batman1 11 0 0 11 Train-only
blacar1 31 0 0 31 Train-only
blctro1 34 0 0 34 Train-only
blgdov1 5 0 0 5 Train-only
blhpar1 86 0 0 86 Train-only
bobfly1 58 0 0 58 Train-only
brratt1 64 0 0 64 Train-only
btfgle1 2 0 0 2 Train-only
cinmou1 89 0 0 89 Train-only
compot1 14 0 0 14 Train-only
duhpar 18 0 0 18 Train-only
eulfly1 3 0 0 3 Train-only
forela1 47 0 0 47 Train-only
garkin1 3 0 0 3 Train-only
gnbtro1 24 0 0 24 Train-only
gogwoo1 10 0 0 10 Train-only
gramou1 80 0 0 80 Train-only
grcfly1 8 0 0 8 Train-only
gretin1 46 0 0 46 Train-only
gycfly1 6 0 0 6 Train-only
gycwor1 48 0 0 48 Train-only
letbar1 73 0 0 73 Train-only
litwoo2 11 0 0 11 Train-only
muswre2 23 0 0 23 Train-only
olioro1 36 0 0 36 Train-only
oliwoo1 55 0 0 55 Train-only
pavpig2 91 0 0 91 Train-only
plbwoo1 26 0 0 26 Train-only
pltant1 57 0 0 57 Train-only
puteup1 28 0 0 28 Train-only
rcatan1 4 0 0 4 Train-only
renwoo1 71 0 0 71 Train-only
rinkin1 12 0 0 12 Train-only
royfly1 2 0 0 2 Train-only
rucant2 13 0 0 13 Train-only
ruftof1 8 0 0 8 Train-only
scapig2 15 0 0 15 Train-only
scbwoo5 21 0 0 21 Train-only
specha3 38 0 0 38 Train-only
squcuc1 7 0 0 7 Train-only
stbwoo2 71 0 0 71 Train-only
strxen1 3 0 0 3 Train-only
stwqua1 27 0 0 27 Train-only
whcspa1 63 0 0 63 Train-only
whfant2 34 0 0 34 Train-only
whltyr1 21 0 0 21 Train-only
whtwoo2 45 0 0 45 Train-only
yectyr1 2 0 0 2 Train-only
yetwoo2 7 0 0 7 Train-only