Datasets:
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
@@ -1,4 +1,11 @@
|
|
1 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
configs:
|
3 |
- config_name: default
|
4 |
data_files:
|
@@ -15,25 +22,30 @@ dataset_info:
|
|
15 |
audio:
|
16 |
sampling_rate: 16000
|
17 |
- name: digit
|
18 |
-
dtype:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
19 |
- name: gender
|
20 |
dtype: string
|
21 |
splits:
|
22 |
- name: train
|
23 |
-
num_bytes:
|
24 |
num_examples: 24000
|
25 |
- name: test
|
26 |
-
num_bytes:
|
27 |
num_examples: 6000
|
28 |
-
download_size:
|
29 |
-
dataset_size:
|
30 |
-
license: mit
|
31 |
-
task_categories:
|
32 |
-
- audio-classification
|
33 |
-
language:
|
34 |
-
- en
|
35 |
-
size_categories:
|
36 |
-
- 10K<n<100K
|
37 |
---
|
38 |
# Dataset Card for "AudioMNIST"
|
39 |
This is just a reupload of the [audioMNIST](https://github.com/soerenab/AudioMNIST) dataset for convenient use with the datasets library. For each speaker (12 female, 48 male), there are 50 recordings of each digit.
|
|
|
1 |
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
license: mit
|
5 |
+
size_categories:
|
6 |
+
- 10K<n<100K
|
7 |
+
task_categories:
|
8 |
+
- audio-classification
|
9 |
configs:
|
10 |
- config_name: default
|
11 |
data_files:
|
|
|
22 |
audio:
|
23 |
sampling_rate: 16000
|
24 |
- name: digit
|
25 |
+
dtype:
|
26 |
+
class_label:
|
27 |
+
names:
|
28 |
+
'0': '0'
|
29 |
+
'1': '1'
|
30 |
+
'2': '2'
|
31 |
+
'3': '3'
|
32 |
+
'4': '4'
|
33 |
+
'5': '5'
|
34 |
+
'6': '6'
|
35 |
+
'7': '7'
|
36 |
+
'8': '8'
|
37 |
+
'9': '9'
|
38 |
- name: gender
|
39 |
dtype: string
|
40 |
splits:
|
41 |
- name: train
|
42 |
+
num_bytes: 1492088102.0
|
43 |
num_examples: 24000
|
44 |
- name: test
|
45 |
+
num_bytes: 360692430.0
|
46 |
num_examples: 6000
|
47 |
+
download_size: 1483317200
|
48 |
+
dataset_size: 1852780532.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
49 |
---
|
50 |
# Dataset Card for "AudioMNIST"
|
51 |
This is just a reupload of the [audioMNIST](https://github.com/soerenab/AudioMNIST) dataset for convenient use with the datasets library. For each speaker (12 female, 48 male), there are 50 recordings of each digit.
|