Datasets:
Tasks:
Automatic Speech Recognition
Formats:
parquet
Languages:
Danish
Size:
1M - 10M
ArXiv:
License:
Commit
·
343e05c
1
Parent(s):
2ab4865
Update README.md
Browse files
README.md
CHANGED
@@ -114,7 +114,7 @@ There are 57 unique speakers in the training dataset and 25 unique speakers in t
|
|
114 |
|
115 |
#### Transcription Length Distribution
|
116 |
|
117 |
-

|
118 |
|
119 |
## Dataset Creation
|
120 |
|