Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,162 +1,97 @@
|
|
| 1 |
---
|
| 2 |
-
license:
|
| 3 |
-
dataset_info:
|
| 4 |
-
features:
|
| 5 |
-
- name: audio
|
| 6 |
-
dtype: audio
|
| 7 |
-
splits:
|
| 8 |
-
- name: train
|
| 9 |
-
num_bytes: 20465198175
|
| 10 |
-
num_examples: 15814
|
| 11 |
-
download_size: 18305122770
|
| 12 |
-
dataset_size: 20465198175
|
| 13 |
-
configs:
|
| 14 |
-
- config_name: default
|
| 15 |
-
data_files:
|
| 16 |
-
- split: train
|
| 17 |
-
path: data/train-*
|
| 18 |
-
task_categories:
|
| 19 |
-
- automatic-speech-recognition
|
| 20 |
-
language:
|
| 21 |
-
- vi
|
| 22 |
-
tags:
|
| 23 |
-
- medical
|
| 24 |
-
- code-switching
|
| 25 |
---
|
| 26 |
-
#
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
-
|
| 43 |
-
-
|
| 44 |
-
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
|
| 53 |
-
|
| 54 |
-
|
| 55 |
-
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
###
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
|
| 80 |
-
|
| 81 |
-
|
| 82 |
-
|
| 83 |
-
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
| 90 |
-
|
| 91 |
-
|
| 92 |
-
|
| 93 |
-
|
| 94 |
-
|
| 95 |
-
|
| 96 |
-
|
| 97 |
-
|
| 98 |
-
|
| 99 |
-
|
| 100 |
-
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
|
| 106 |
-
|
| 107 |
-
|
| 108 |
-
|
| 109 |
-
|
| 110 |
-
|
| 111 |
-
|
| 112 |
-
|
| 113 |
-
|
| 114 |
-
|
| 115 |
-
|
| 116 |
-
|
| 117 |
-
|
| 118 |
-
|
| 119 |
-
|
| 120 |
-
[More Information Needed]
|
| 121 |
-
|
| 122 |
-
## Bias, Risks, and Limitations
|
| 123 |
-
|
| 124 |
-
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 125 |
-
|
| 126 |
-
[More Information Needed]
|
| 127 |
-
|
| 128 |
-
### Recommendations
|
| 129 |
-
|
| 130 |
-
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 131 |
-
|
| 132 |
-
Users should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.
|
| 133 |
-
|
| 134 |
-
## Citation [optional]
|
| 135 |
-
|
| 136 |
-
<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->
|
| 137 |
-
|
| 138 |
-
**BibTeX:**
|
| 139 |
-
|
| 140 |
-
[More Information Needed]
|
| 141 |
-
|
| 142 |
-
**APA:**
|
| 143 |
-
|
| 144 |
-
[More Information Needed]
|
| 145 |
-
|
| 146 |
-
## Glossary [optional]
|
| 147 |
-
|
| 148 |
-
<!-- If relevant, include terms and calculations in this section that can help readers understand the dataset or dataset card. -->
|
| 149 |
-
|
| 150 |
-
[More Information Needed]
|
| 151 |
-
|
| 152 |
-
## More Information [optional]
|
| 153 |
-
|
| 154 |
-
[More Information Needed]
|
| 155 |
-
|
| 156 |
-
## Dataset Card Authors [optional]
|
| 157 |
-
|
| 158 |
-
[More Information Needed]
|
| 159 |
-
|
| 160 |
-
## Dataset Card Contact
|
| 161 |
-
|
| 162 |
-
[More Information Needed]
|
|
|
|
| 1 |
---
|
| 2 |
+
license: other
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
| 4 |
+
# 🩺 ViMedCSS: Vietnamese Medical Code-Switching Speech Dataset
|
| 5 |
+
|
| 6 |
+
## 📖 Overview
|
| 7 |
+
ViMedCSS is a Vietnamese medical speech dataset for code-switching ASR, where each utterance contains at least one non-Vietnamese (mainly English) medical term embedded in Vietnamese speech.
|
| 8 |
+
|
| 9 |
+
This dataset card is prepared from:
|
| 10 |
+
- The metadata files in `ViMedCSS-Metadata` (`train_set.csv`, `valid_set.csv`, `test_set.csv`, `hard_set.csv`)
|
| 11 |
+
- The paper: `Improving_Code_Switching_Detection_of_ASR_Models_for_Medical_Vietnamese_Speech (1).pdf`
|
| 12 |
+
|
| 13 |
+
From the paper, the full ViMedCSS corpus is reported as:
|
| 14 |
+
- 34.57 hours
|
| 15 |
+
- 16,576 utterances
|
| 16 |
+
- 5 medical topics
|
| 17 |
+
- 4 splits including a dedicated `hard` split for rare/unseen terms
|
| 18 |
+
|
| 19 |
+
From the provided CSV metadata in this release, the indexed subset contains:
|
| 20 |
+
- 15,818 rows
|
| 21 |
+
- 15,814 unique `segment_id`
|
| 22 |
+
- 32.64 hours total duration
|
| 23 |
+
- 16,581 total code-switched term occurrences
|
| 24 |
+
- 889 distinct code-switched medical terms
|
| 25 |
+
|
| 26 |
+
## 📊 Dataset Statistics
|
| 27 |
+
### Split Statistics (from `ViMedCSS-Metadata`)
|
| 28 |
+
| Split | # Rows | Duration (hours) | Avg duration (s) | Total CS terms |
|
| 29 |
+
|---|---:|---:|---:|---:|
|
| 30 |
+
| train | 11,832 | 24.30 | 7.39 | 12,314 |
|
| 31 |
+
| validation | 1,714 | 3.57 | 7.49 | 1,814 |
|
| 32 |
+
| test | 1,614 | 3.39 | 7.56 | 1,695 |
|
| 33 |
+
| hard | 658 | 1.38 | 7.57 | 758 |
|
| 34 |
+
| **Total** | **15,818** | **32.64** | **7.43** | **16,581** |
|
| 35 |
+
|
| 36 |
+
### Topic Statistics (from `ViMedCSS-Metadata`)
|
| 37 |
+
| Topic | # Rows | Duration (hours) | Total CS terms |
|
| 38 |
+
|---|---:|---:|---:|
|
| 39 |
+
| Medical Sciences | 6,836 | 14.68 | 7,459 |
|
| 40 |
+
| Pathology & Pathogens | 4,827 | 10.00 | 4,951 |
|
| 41 |
+
| Treatments | 1,969 | 3.80 | 1,985 |
|
| 42 |
+
| Nutrition | 1,155 | 2.14 | 1,155 |
|
| 43 |
+
| Diagnostics | 1,031 | 2.02 | 1,031 |
|
| 44 |
+
|
| 45 |
+
## 🧾 Data Fields
|
| 46 |
+
Each row in metadata corresponds to one segment audio file, where:
|
| 47 |
+
- `segment_id` maps to `segment_id.wav` (for example: `Med_CS-100-17` -> `Med_CS-100-17.wav`)
|
| 48 |
+
|
| 49 |
+
Main fields:
|
| 50 |
+
- `segment_id`: utterance identifier
|
| 51 |
+
- `duration_seconds`: utterance duration
|
| 52 |
+
- `segment_text`: Vietnamese transcript containing code-switched term(s)
|
| 53 |
+
- `cs_terms_list`: semicolon-separated code-switched terms
|
| 54 |
+
- `cs_terms_count`: number of code-switched terms in the utterance
|
| 55 |
+
- `topic` (or `Topic` in one CSV): medical topic label
|
| 56 |
+
- `original_video_link`: source video URL
|
| 57 |
+
- `original_video_title`: source video title
|
| 58 |
+
- `start_time`, `end_time`: segment boundaries in source audio/video
|
| 59 |
+
|
| 60 |
+
When loaded from Hugging Face, an `audio` column is available with waveform bytes/path in the standard 🤗 Datasets `Audio` format.
|
| 61 |
+
|
| 62 |
+
## 🔽 How to Load
|
| 63 |
+
Load directly with 🤗 Datasets:
|
| 64 |
+
|
| 65 |
+
```python
|
| 66 |
+
from datasets import load_dataset
|
| 67 |
+
|
| 68 |
+
dataset = load_dataset("tensorxt/ViMedCSS")
|
| 69 |
+
print(dataset)
|
| 70 |
+
```
|
| 71 |
+
|
| 72 |
+
Clone with Git LFS:
|
| 73 |
+
|
| 74 |
+
```bash
|
| 75 |
+
git lfs install
|
| 76 |
+
git clone https://huggingface.co/datasets/tensorxt/ViMedCSS
|
| 77 |
+
```
|
| 78 |
+
|
| 79 |
+
## 📝 Notes
|
| 80 |
+
- The paper reports the full corpus statistics (34.57h, 16,576 utterances).
|
| 81 |
+
- The CSV metadata bundled here describes a smaller processed subset (32.64h, 15,818 rows).
|
| 82 |
+
- The `hard` split is intended for evaluating rare/unseen code-switched medical terms, following the paper’s benchmark setup.
|
| 83 |
+
|
| 84 |
+
## 📜 License
|
| 85 |
+
The paper states that data are collected from publicly available YouTube content for research purposes, and the medical dictionary resource used in construction is under institutional intellectual property licensing.
|
| 86 |
+
|
| 87 |
+
Please verify usage rights for your setting before redistribution or commercial use.
|
| 88 |
+
|
| 89 |
+
## 🙏 Citation
|
| 90 |
+
If you use ViMedCSS, please cite:
|
| 91 |
+
|
| 92 |
+
```bibtex
|
| 93 |
+
@misc{nguyen_vimedcss,
|
| 94 |
+
title={ViMedCSS: A Vietnamese Medical Code-Switching Speech Dataset \& Benchmark},
|
| 95 |
+
author={Nguyen, Tung X. and Vo, Nhu and Nguyen, Giang-Son and Hoang, Duy Mai and Huynh, Chien Dinh and Jauregi Unanue, Inigo and Piccardi, Massimo and Buntine, Wray and Le, Dung D.},
|
| 96 |
+
}
|
| 97 |
+
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|