Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -36,6 +36,9 @@ size_categories:
|
|
| 36 |
- 10K<n<100K
|
| 37 |
---
|
| 38 |
# Dataset Card for "AudioMNIST"
|
|
|
|
|
|
|
|
|
|
| 39 |
|
| 40 |
### Citation Information
|
| 41 |
The original creators of the dataset ask you to cite [their paper](https://arxiv.org/abs/1807.03418) if you use this data:
|
|
|
|
| 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.
|
| 40 |
+
I created an 80/20 train test split, making sure that no speaker is contained in both the train and the test.
|
| 41 |
+
|
| 42 |
|
| 43 |
### Citation Information
|
| 44 |
The original creators of the dataset ask you to cite [their paper](https://arxiv.org/abs/1807.03418) if you use this data:
|