Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -48,10 +48,15 @@ The raw data collection has over 600,000 data instances (375 hours). After proce
|
|
48 |
- WAV (path to audio file, located in the data directory)
|
49 |
- EMOTION (annotated emotion)
|
50 |
- INTENSITY (annotated intensity [1-5], 1 corresponds to low intensity and 5 corresponds to high intensity, 0 corresponds to no annotation)
|
|
|
51 |
|
52 |
### Data Splits
|
53 |
|
54 |
-
|
|
|
|
|
|
|
|
|
55 |
|
56 |
## Dataset Creation
|
57 |
|
|
|
48 |
- WAV (path to audio file, located in the data directory)
|
49 |
- EMOTION (annotated emotion)
|
50 |
- INTENSITY (annotated intensity [1-5], 1 corresponds to low intensity and 5 corresponds to high intensity, 0 corresponds to no annotation)
|
51 |
+
- LENGTH
|
52 |
|
53 |
### Data Splits
|
54 |
|
55 |
+
The data is split into train, test, validation sets (located in the corresponding json manifest files).
|
56 |
+
- Train [80%]
|
57 |
+
- Validation [10%]
|
58 |
+
- Test [10%]
|
59 |
+
The unsplit data is also available in data.csv, in order to add flexibility with custom splits.
|
60 |
|
61 |
## Dataset Creation
|
62 |
|