Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,155 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
task_categories:
|
| 3 |
+
- audio-classification
|
| 4 |
+
---
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
### Datasets
|
| 8 |
+
|
| 9 |
+
#### Train
|
| 10 |
+
- Exclusively using focal audio data from Xeno-Canto (XC) with quality ratings A, B, C and excluding all recordings that are CC-ND.
|
| 11 |
+
- Each dataset is tailored for specific target species identified in soundscape files.
|
| 12 |
+
- We offer detected events and corresponding cluster assignments to identify bird sounds in each recording.
|
| 13 |
+
- We provide the full recordings from XC! These can generate multiple samples from a single instance.
|
| 14 |
+
|
| 15 |
+
#### Test
|
| 16 |
+
- Only soundscape data sourced from Zenodo.
|
| 17 |
+
- We provide the full recording with the complete label set and specified bounding boxes.
|
| 18 |
+
- This dataset excludes recordings that do not contain bird calls ("no_call").
|
| 19 |
+
- Task: Multiclass ("ebird_code")
|
| 20 |
+
|
| 21 |
+
#### Test_5s
|
| 22 |
+
- Only soundscape data from Zenodo formatted acoording to the Kaggle evaluation scheme.
|
| 23 |
+
- Each recording is segmented into 5-second intervals without overlaps.
|
| 24 |
+
- This contains the "no_call" class.
|
| 25 |
+
- Task: Multilabel ("ebird_code_multilabel")
|
| 26 |
+
|
| 27 |
+
### Subsets
|
| 28 |
+
|
| 29 |
+
Numbers need to be updated
|
| 30 |
+
|
| 31 |
+
| | train | test | test_5s | size (GB) | #classes |
|
| 32 |
+
|----------------------------|--------:|-----------:|--------:|-----------:|-------------:|
|
| 33 |
+
| PER (amazon_basin) | 21,834 | 14,798 | 15,120 | 13 | 132 |
|
| 34 |
+
| NES (columbia_costa_rica) | 4,650 | 6,952 | 24,480 | 17 | 89 |
|
| 35 |
+
| UHH (hawaiian_islands) | 4,611 | 59,583 | 36,637 | 6.3 | 25 tr, 27 te |
|
| 36 |
+
| HSN (high_sierras) | 6,526 | 10,296 | 12,000 | 6.9 | 21 |
|
| 37 |
+
| NBP (nips) | 100,174| 5,493 | 563 | 162 | 51 tr, 52 te|
|
| 38 |
+
| POW (powdermill_nature) | 17,059 | 16,052 | 4560| 18 | 48 |
|
| 39 |
+
| SSW (sapsucker_woods) | 32,902 | 50,760 | 205,200| 40 | 81 |
|
| 40 |
+
| SNE (sierra_nevada) | 23,164 | 20,147 | 23,756| 25 | 56 |
|
| 41 |
+
| XCL (xenocanto) | 686,593| x | | 648 | 10,124 |
|
| 42 |
+
| XCM | 80,012 | x | | | 410 |
|
| 43 |
+
|
| 44 |
+
#### FEATURES
|
| 45 |
+
|
| 46 |
+
```python
|
| 47 |
+
{
|
| 48 |
+
"audio": datasets.Audio(sampling_rate=32_000, mono=True, decode=True),
|
| 49 |
+
"filepath": datasets.Value("string"),
|
| 50 |
+
"start_time": datasets.Value("float64"),
|
| 51 |
+
"end_time": datasets.Value("float64"),
|
| 52 |
+
"low_freq": datasets.Value("int64"),
|
| 53 |
+
"high_freq": datasets.Value("int64"),
|
| 54 |
+
"ebird_code": datasets.ClassLabel(names=class_list),
|
| 55 |
+
"ebird_code_multilabel": datasets.Sequence(datasets.ClassLabel(names=["no_call"] + class_list)),
|
| 56 |
+
"ebird_code_secondary": datasets.Sequence(datasets.Value("string")),
|
| 57 |
+
"call_type": datasets.Value("string"),
|
| 58 |
+
"sex": datasets.Value("string"),
|
| 59 |
+
"lat": datasets.Value("float64"),
|
| 60 |
+
"long": datasets.Value("float64"),
|
| 61 |
+
"length": datasets.Value("int64"),
|
| 62 |
+
"microphone": datasets.Value("string"),
|
| 63 |
+
"license": datasets.Value("string"),
|
| 64 |
+
"source": datasets.Value("string"),
|
| 65 |
+
"local_time": datasets.Value("string"),
|
| 66 |
+
"detected_events": datasets.Sequence(datasets.Sequence(datasets.Value("float64"))),
|
| 67 |
+
"event_cluster": datasets.Sequence(datasets.Value("int64")),
|
| 68 |
+
"quality": datasets.Value("string"),
|
| 69 |
+
"recordist": datasets.Value("string")
|
| 70 |
+
})
|
| 71 |
+
```
|
| 72 |
+
```python
|
| 73 |
+
EXAMPLE TRAIN
|
| 74 |
+
{'audio': {'path': '.ogg',
|
| 75 |
+
'array': array([ 6.24680333e-02, 7.57145062e-02, 4.91199419e-02, ...,
|
| 76 |
+
-2.04162002e-02, 8.73558223e-03, -6.23320229e-05]),
|
| 77 |
+
'sampling_rate': 32000},
|
| 78 |
+
'filepath': '.ogg',
|
| 79 |
+
'start_time': None,
|
| 80 |
+
'end_time': None,
|
| 81 |
+
'low_freq': None,
|
| 82 |
+
'high_freq': None,
|
| 83 |
+
'ebird_code': 1,
|
| 84 |
+
'ebird_code_multiclass': None,
|
| 85 |
+
'call_type': 'call',
|
| 86 |
+
'sex': None,
|
| 87 |
+
'lat': 22.2029,
|
| 88 |
+
'long': -159.473,
|
| 89 |
+
'microphone': 'focal',
|
| 90 |
+
'license': '//creativecommons.org/licenses/by-nc-sa/4.0/',
|
| 91 |
+
'source': 'xenocanto',
|
| 92 |
+
'local_time': '12:49',
|
| 93 |
+
'detected_events': [[0.832, 2.48],
|
| 94 |
+
[2.992, 4.016],
|
| 95 |
+
[3.2, 3.904],
|
| 96 |
+
[5.472, 6.048],
|
| 97 |
+
[5.488, 6.432],
|
| 98 |
+
[7.088, 8.16],
|
| 99 |
+
[8.944, 10.432],
|
| 100 |
+
[10.72, 12.672],
|
| 101 |
+
[11.152, 13.2],
|
| 102 |
+
[13.488, 14.0],
|
| 103 |
+
[14.64, 16.496]],
|
| 104 |
+
'event_cluster': [1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1],
|
| 105 |
+
'quality': 'A'}
|
| 106 |
+
|
| 107 |
+
EXAMPLE TEST_5S
|
| 108 |
+
{'audio': {'path': '.ogg',
|
| 109 |
+
'array': array([-5.03219722e-04, 9.99580720e-04, 2.58744985e-05, ...,
|
| 110 |
+
-4.06746846e-03, -3.79991997e-03, 2.88472045e-04]),
|
| 111 |
+
'sampling_rate': 32000},
|
| 112 |
+
'filepath': '.ogg',
|
| 113 |
+
'start_time': 0.0,
|
| 114 |
+
'end_time': 5.0,
|
| 115 |
+
'low_freq': 2678,
|
| 116 |
+
'high_freq': 6053,
|
| 117 |
+
'ebird_code': None,
|
| 118 |
+
'ebird_code_multiclass': [0],
|
| 119 |
+
'call_type': None,
|
| 120 |
+
'sex': None,
|
| 121 |
+
'lat': 19.801668,
|
| 122 |
+
'long': -155.609444,
|
| 123 |
+
'microphone': 'Soundscape',
|
| 124 |
+
'license': 'Creative Commons Attribution 4.0 International Public License',
|
| 125 |
+
'source': 'https://zenodo.org/record/7078499',
|
| 126 |
+
'local_time': '15:00:06',
|
| 127 |
+
'detected_events': None,
|
| 128 |
+
'event_cluster': None,
|
| 129 |
+
'quality': None}
|
| 130 |
+
|
| 131 |
+
EXAMPLE TEST
|
| 132 |
+
{'audio': {'path': '.ogg',
|
| 133 |
+
'array': array([-5.03219722e-04, 9.99580720e-04, 2.58744985e-05, ...,
|
| 134 |
+
-4.06746846e-03, -3.79991997e-03, 2.88472045e-04]),
|
| 135 |
+
'sampling_rate': 32000},
|
| 136 |
+
'filepath': '.ogg',
|
| 137 |
+
'start_time': 6.8,
|
| 138 |
+
'end_time': 8.2,
|
| 139 |
+
'low_freq': 2678,
|
| 140 |
+
'high_freq': 6053,
|
| 141 |
+
'ebird_code': 22,
|
| 142 |
+
'ebird_code_multiclass': None,
|
| 143 |
+
'call_type': None,
|
| 144 |
+
'sex': None,
|
| 145 |
+
'lat': 19.801668,
|
| 146 |
+
'long': -155.609444,
|
| 147 |
+
'microphone': 'Soundscape',
|
| 148 |
+
'license': 'Creative Commons Attribution 4.0 International Public License',
|
| 149 |
+
'source': 'https://zenodo.org/record/7078499',
|
| 150 |
+
'local_time': '15:00:06',
|
| 151 |
+
'detected_events': None,
|
| 152 |
+
'event_cluster': None,
|
| 153 |
+
'quality': None}
|
| 154 |
+
```
|
| 155 |
+
|