Taejin commited on
Commit
c45cc76
1 Parent(s): f2c334a

Update the new v2 split

Browse files

Updating the split.
Created training split
and reduced the dev and eval sessions.

Files changed (1) hide show
  1. README.md +13 -7
README.md CHANGED
@@ -42,12 +42,15 @@ Example) `session_ge1nse2c.seglst.json`
42
 
43
  The dataset is divided into training and test splits:
44
 
45
- - Development Data: 142 entries
46
- - 2 to 4 speakers in each session
47
- - Approximately 10 ~ 40 mins of recordings
48
- - Evaluation Data: 104 entries
49
  - 2 to 4 speakers in each session
50
  - Approximately 10 ~ 40 mins of recordings
 
 
 
 
 
 
51
 
52
  ### Keys (items)
53
 
@@ -62,15 +65,18 @@ The dataset is divided into training and test splits:
62
  `err_source_text`: This is the erroneous ASR-Diarization results to be fixed. Has dev, eval folders
63
  `ref_annotated_text`: This is the human annotated ground-truth for evaluation. Only dev split is included.
64
 
 
 
 
65
  - **Development Sources**:
66
- - `dev`: 142 sessions
67
 
68
  - **Evaluation Sources**:
69
- - `eval`: 104 Sessions
70
 
71
  ## Access
72
 
73
- The dataset can be accessed and downloaded through the HuggingFace Datasets library.
74
 
75
  ## Evaluation
76
 
 
42
 
43
  The dataset is divided into training and test splits:
44
 
45
+ - Training Data: 222 entries
 
 
 
46
  - 2 to 4 speakers in each session
47
  - Approximately 10 ~ 40 mins of recordings
48
+ - Development Data: 13 entries
49
+ - 2 speakers in each session
50
+ - Approximately 10 mins of recordings
51
+ - Evaluation Data: 11 entries
52
+ - 2 speakers in each session
53
+ - Approximately 10 mins of recordings
54
 
55
  ### Keys (items)
56
 
 
65
  `err_source_text`: This is the erroneous ASR-Diarization results to be fixed. Has dev, eval folders
66
  `ref_annotated_text`: This is the human annotated ground-truth for evaluation. Only dev split is included.
67
 
68
+ - **Training Sources**:
69
+ - `dev`: 222 sessions
70
+
71
  - **Development Sources**:
72
+ - `dev`: 13 sessions
73
 
74
  - **Evaluation Sources**:
75
+ - `eval`: 11 Sessions
76
 
77
  ## Access
78
 
79
+ The dataset can be accessed and downloaded through the HuggingFace Datasets library (i.e., This Repository).
80
 
81
  ## Evaluation
82