Datasets:

disco-eth
/

EuroSpeech

Tasks:

Automatic Speech Recognition

Modalities:

Formats:

Languages:

Size:

Libraries:

License:

Dataset card Data Studio Files Files and versions

Samuel Pfisterer commited on May 17, 2025

Commit

4bbc219

·

verified ·

1 Parent(s): bbf5f7e

Update README.md

Files changed (1) hide show

README.md +1 -1

README.md CHANGED Viewed

@@ -1076,7 +1076,7 @@ EuroSpeech provides substantial data for previously under-resourced languages:
 - 19 languages exceed 1,000 hours of data (CER < 20%)
 - 22 languages exceed 500 hours of data (CER < 20%)
-| Language         | Code & Total Aligned (h) & CER < 30\% (h) & CER < 20\% (h) & CER < 10\% (h) |
 |------------------|------|-------------------|----------------|----------------|----------------|
 | Croatia          | hr   | 7484.9            | 5899.7         | 5615.8         | 4592.0         |
 | Denmark          | da   | 7014.2            | 6435.0         | 5559.8         | 3443.7         |

 - 19 languages exceed 1,000 hours of data (CER < 20%)
 - 22 languages exceed 500 hours of data (CER < 20%)
+| Language         | Code | Total Aligned (h) | CER < 30\% (h) | CER < 20\% (h) | CER < 10\% (h) |
 |------------------|------|-------------------|----------------|----------------|----------------|
 | Croatia          | hr   | 7484.9            | 5899.7         | 5615.8         | 4592.0         |
 | Denmark          | da   | 7014.2            | 6435.0         | 5559.8         | 3443.7         |