Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -7,6 +7,19 @@ configs:
|
|
7 |
- split: train
|
8 |
path: "data/train.parquet"
|
9 |
---
|
|
|
10 |
It is a dataset of ukrainian audiobooks
|
11 |
-
Each sample contain an approximately 8
|
12 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
- split: train
|
8 |
path: "data/train.parquet"
|
9 |
---
|
10 |
+
### About dataset
|
11 |
It is a dataset of ukrainian audiobooks
|
12 |
+
Each sample contain an approximately 8 seconds od ukrainian speech
|
13 |
|
14 |
+
### Loading script
|
15 |
+
'''
|
16 |
+
>>> load_dataset("Zarakun/audiobooks_ua_test")
|
17 |
+
'''
|
18 |
+
|
19 |
+
### Dataset structure
|
20 |
+
**Every example has the following:
|
21 |
+
**audio** - the waveform
|
22 |
+
**rate** - the sampling rate of the waveform
|
23 |
+
**file_id** - the id of the speaker
|
24 |
+
**duration** - the duration of the video in seconds
|
25 |
+
**sentence** - the transcript of the video
|