ASLP-lab commited on
Commit
cb305bf
·
verified ·
1 Parent(s): b63ac30

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +9 -9
README.md CHANGED
@@ -1,21 +1,21 @@
1
  ---
2
  license: apache-2.0
3
  ---
4
- ## 📊 WSYue-ASR-eval: Cantonese ASR Benchmark
5
 
6
  To address the unique linguistic characteristics of Cantonese, we propose **WSYue-eval**, a comprehensive benchmark encompassing both **ASR** and **TTS** tasks. This integrated evaluation framework is specifically tailored to assess model performance across critical dimensions of Cantonese language processing.
7
 
8
- ### 🔹 ASR Benchmark
9
- As a representative task of speech understanding, we developed the **WSYue-ASR-eval** test set.
10
- - Annotated through multiple rounds of **manual labeling**
11
- - Includes rich tags such as **text transcription, emotion, age, and gender**
12
- - Covers **Cantonese-English code-switching** and **multi-domain conditions**
13
- - Enables comprehensive evaluation across varying speech lengths
14
 
15
- ### 📑 WSYue-ASR-eval Subsets
16
  | Set | Duration | Speakers | Hours |
17
  |-------|----------|----------|-------|
18
  | Short | 0–10 s | 2861 | 9.46 |
19
  | Long | 10–30 s | 838 | 1.97 |
20
 
21
- Total: **~11.4 hours**, with diverse speakers and scenarios.
 
1
  ---
2
  license: apache-2.0
3
  ---
4
+ # WSYue-ASR-eval: Cantonese ASR Benchmark
5
 
6
  To address the unique linguistic characteristics of Cantonese, we propose **WSYue-eval**, a comprehensive benchmark encompassing both **ASR** and **TTS** tasks. This integrated evaluation framework is specifically tailored to assess model performance across critical dimensions of Cantonese language processing.
7
 
8
+ ## ASR Benchmark
9
+ As a representative task of speech understanding, we developed the **WSYue-ASR-eval** test set for the ASR task.
10
+ - Annotated through multiple rounds of manual labeling.
11
+ - Includes rich tags such as text transcription, emotion, age, and gender.
12
+ - Covers Cantonese-English code-switching and multi-domain conditions.
13
+ - Enables comprehensive evaluation across varying speech lengths.
14
 
15
+ ## WSYue-ASR-eval Subsets
16
  | Set | Duration | Speakers | Hours |
17
  |-------|----------|----------|-------|
18
  | Short | 0–10 s | 2861 | 9.46 |
19
  | Long | 10–30 s | 838 | 1.97 |
20
 
21
+ Total: 11.4 hours, with diverse speakers and scenarios.