Skip to content

Releases: Alyetama/BirdFSD-YOLOv5

v1.0.0-alpha.6

15 Jun 23:49
Compare
Choose a tag to compare
v1.0.0-alpha.6 Pre-release
Pre-release

Training start time: 2022-06-14T16:25:02
Training duration: 04:22:31
W&B run URL: https://wandb.ai/biodiv/train/runs/y31qiwv2
W&B run ID: y31qiwv2
W&B run name: proud-glade-102
W&B run path: biodiv/train/y31qiwv2
Number of classes: 15
Dataset name: dataset-YOLO-06-13-2022_19.55.50.tar

System hardware
CPU count GPU count GPU type NVIDIA driver version
48 2 Tesla T4, Tesla T4 470.103.01
Base ML framework
Python CUDA Torch Torchvision
3.8.12 10.2 1.10.0 0.11.0a0+fb5bd47
Classes
{
    "American Goldfinch (Spinus tristis)": 136,
    "Brown-headed Cowbird (Molothrus ater)": 186,
    "Carolina Chickadee (Poecile carolinensis)": 335,
    "Chipping Sparrow (Spizella passerina)": 1061,
    "Downy Woodpecker (Dryobates pubescens)": 170,
    "Eastern Towhee (Pipilo erythrophthalmus)": 111,
    "Hermit Thrush (Catharus guttatus)": 141,
    "House Finch (Haemorhous mexicanus)": 643,
    "Northern Cardinal (Cardinalis cardinalis)": 547,
    "Red-bellied Woodpecker (Melanerpes carolinus)": 32,
    "Song Sparrow (Melospiza melodia)": 126,
    "Squirrels (Eutamias sibiricus)": 1817,
    "Tufted Titmouse (Baeolophus bicolor)": 459,
    "White-throated Sparrow (Zonotrichia albicollis)": 32,
    "Yellow-rumped Warbler (Setophaga coronata)": 405
}
classes-hist
Validation predictions val_batch1_labels-BirdFSD-YOLOv5-v1.0.0-alpha.6.jpg val_batch2_pred-BirdFSD-YOLOv5-v1.0.0-alpha.6.jpg val_batch0_labels-BirdFSD-YOLOv5-v1.0.0-alpha.6.jpg val_batch1_pred-BirdFSD-YOLOv5-v1.0.0-alpha.6.jpg val_batch0_pred-BirdFSD-YOLOv5-v1.0.0-alpha.6.jpg val_batch2_labels-BirdFSD-YOLOv5-v1.0.0-alpha.6.jpg
Confusion matrix cm

Model performance:

best/precision best/recall val/box_loss val/cls_loss val/obj_loss F1-score
0.9818 0.9735 0.0162 0.0034 0.0055 0.9776

v1.0.0-alpha.5

03 Jun 02:39
Compare
Choose a tag to compare
v1.0.0-alpha.5 Pre-release
Pre-release

Training start time: 2022-06-02T14:24:29
Training duration: 08:40:50
W&B run URL: https://wandb.ai/biodiv/train/runs/1z45wzvc
W&B run ID: 1z45wzvc
W&B run name: efficient-waterfall-96
W&B run path: biodiv/train/1z45wzvc
Number of classes: 14
Dataset name: dataset-YOLO-06-02-2022_09.10.43.tar

System hardware
CPU count GPU count GPU type NVIDIA driver version
40 2 Tesla V100-PCIE-32GB, Tesla V100-PCIE-32GB 470.103.01
Base ML framework
Python CUDA Torch Torchvision
3.8.13 10.2 1.11.0 0.12.0
Classes
{
    "American Goldfinch (Spinus tristis)": 130,
    "Brown-headed Cowbird (Molothrus ater)": 186,
    "Carolina Chickadee (Poecile carolinensis)": 332,
    "Chipping Sparrow (Spizella passerina)": 1061,
    "Downy Woodpecker (Dryobates pubescens)": 122,
    "Eastern Towhee (Pipilo erythrophthalmus)": 111,
    "Hermit Thrush (Catharus guttatus)": 141,
    "House Finch (Haemorhous mexicanus)": 611,
    "Northern Cardinal (Cardinalis cardinalis)": 547,
    "Song Sparrow (Melospiza melodia)": 126,
    "Squirrels (Eutamias sibiricus)": 296,
    "Tufted Titmouse (Baeolophus bicolor)": 431,
    "White-throated Sparrow (Zonotrichia albicollis)": 32,
    "Yellow-rumped Warbler (Setophaga coronata)": 405
}
classes-hist
Validation predictions val_batch1_labels-BirdFSD-YOLOv5-v1.0.0-alpha.5.jpg val_batch2_pred-BirdFSD-YOLOv5-v1.0.0-alpha.5.jpg val_batch0_labels-BirdFSD-YOLOv5-v1.0.0-alpha.5.jpg val_batch1_pred-BirdFSD-YOLOv5-v1.0.0-alpha.5.jpg val_batch0_pred-BirdFSD-YOLOv5-v1.0.0-alpha.5.jpg val_batch2_labels-BirdFSD-YOLOv5-v1.0.0-alpha.5.jpg
Confusion matrix cm

Model performance:

best/precision best/recall val/box_loss val/cls_loss val/obj_loss F1-score
0.9797 0.8992 0.0162 0.0025 0.006 0.9377

v1.0.0-alpha.4

07 May 07:43
Compare
Choose a tag to compare
v1.0.0-alpha.4 Pre-release
Pre-release

DEPRECATED.


Training start time: 2022-05-06T05:38:22
Training duration: 19:59:24
W&B run URL: https://wandb.ai/biodiv/train/runs/808p95Vj
W&B run ID: 808p95Vj
W&B run name: peach-wave-60
W&B run path: biodiv/train/808p95Vj
Number of classes: 14
Dataset name: dataset-YOLO-05-06-2022_02.57.11.tar

System hardware
CPU count GPU count GPU type NVIDIA driver version
40 1 Tesla V100-PCIE-32GB 470.103.01
Base ML framework
Python CUDA Torch Torchvision
3.8.13 10.2 1.11.0 0.12.0
Classes
{
    "American Goldfinch": 118,
    "Brown-headed Cowbird": 186,
    "Carolina Chickadee": 324,
    "Chipping Sparrow": 1084,
    "Downy Woodpecker": 114,
    "Hermit Thrush": 141,
    "House Finch": 566,
    "Northern Cardinal": 547,
    "Song Sparrow": 53,
    "Spotted/Eastern Towhee (Rufous-sided Towhee)": 110,
    "Tufted Titmouse": 418,
    "White-throated Sparrow": 32,
    "Yellow-rumped Warbler": 404,
    "squirrels": 248
}
Validation predictions val-0 val-1 val-2 val-3 val-4 val-5
Confusion matrix cm

Model performance:

best/precision best/recall val/box_loss val/cls_loss val/obj_loss F1-score
0.9509 0.9908 0.0142 0.0007 0.0067 0.9704

v1.0.0-alpha.3

05 May 21:16
Compare
Choose a tag to compare
v1.0.0-alpha.3 Pre-release
Pre-release

DEPRECATED.


Training start time: 2022-05-05T06:16:44
Training duration: 07:08:22
W&B run URL: https://wandb.ai/biodiv/train/runs/1kwzca8y
W&B run ID: 1kwzca8y
W&B run name: grievous-nexu-55
W&B run path: biodiv/train/1kwzca8y
Number of classes: 21
Dataset name: dataset-YOLO-05-05-2022_00.52.52.tar

System hardware
CPU count GPU count GPU type NVIDIA driver version
48 1 Tesla V100S-PCIE-32GB 470.103.01
Base ML framework
Python CUDA Torch Torchvision
3.8.13 10.2 1.11.0 0.12.0
Classes
{
    "American Goldfinch": 118,
    "American Robin": 9,
    "Brown-headed Cowbird": 186,
    "Brown-headed Nuthatch": 2,
    "Carolina Chickadee": 324,
    "Carolina Wren": 4,
    "Chipping Sparrow": 1084,
    "Dark-eyed Junco": 11,
    "Downy Woodpecker": 114,
    "Eastern Bluebird": 5,
    "Hermit Thrush": 141,
    "House Finch": 566,
    "Mourning Dove": 3,
    "Northern Cardinal": 547,
    "Red-bellied Woodpecker": 3,
    "Song Sparrow": 53,
    "Spotted/Eastern Towhee (Rufous-sided Towhee)": 110,
    "Tufted Titmouse": 418,
    "White-throated Sparrow": 32,
    "Yellow-rumped Warbler": 404,
    "squirrels": 248
}
Validation predictions val-0 val-1 val-2 val-3 val-4 val-5
Confusion matrix cm

Model performance:

best/precision best/recall val/box_loss val/cls_loss val/obj_loss F1-score
0.9029 0.9351 0.0143 0.0015 0.005 0.9187

v1.0.0-alpha.2

22 Apr 00:08
Compare
Choose a tag to compare
v1.0.0-alpha.2 Pre-release
Pre-release

DEPRECATED.


Training start time: 2022-04-21T16:41:48
Training duration: 04:43:27
W&B run URL: https://wandb.ai/biodiv/train/runs/1i9hbms9
W&B run ID: 1i9hbms9
W&B run name: dainty-pyramid-46
W&B run path: biodiv/train/1i9hbms9
Number of classes: 21

Classes
{
    "House Finch": 512,
    "Tufted Titmouse": 484,
    "Northern Cardinal": 492,
    "Brown-headed Cowbird": 37,
    "Carolina Wren": 3,
    "American Robin": 9,
    "Hermit Thrush": 12,
    "Spotted/Eastern Towhee (Rufous-sided Towhee)": 103,
    "American Goldfinch": 119,
    "Chipping Sparrow": 1178,
    "Dark-eyed Junco": 13,
    "Carolina Chickadee": 381,
    "White-throated Sparrow": 2,
    "Song Sparrow": 46,
    "squirrels": 239,
    "Yellow-rumped Warbler": 366,
    "Eastern Bluebird": 2,
    "Downy Woodpecker": 17,
    "Mourning Dove": 3,
    "Brown-headed Nuthatch": 1,
    "Red-bellied Woodpecker": 3
}
Validation predictions (examples) val-0 val-1 val-2 val-3 val-4 val-5
Confusion matrix cm

Model performance:

best/precision best/recall val/box_loss val/cls_loss val/obj_loss F1-score
0.9855 0.8718 0.0157 0.0027 0.0059 0.9251