Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -17,8 +17,6 @@ size_categories:
|
|
| 17 |
|
| 18 |
**HVU_QA** is an open-source Vietnamese Question–Context–Answer (QCA) corpus and supporting tools for building FAQ-style question generation systems in low-resource languages. The dataset was created using a fully automated pipeline that combines web crawling from trustworthy sources, semantic tag-based extraction, and AI-assisted filtering to ensure high factual accuracy.
|
| 19 |
|
| 20 |
-
---
|
| 21 |
-
|
| 22 |
## 📋 Dataset Description
|
| 23 |
|
| 24 |
- **Language:** Vietnamese
|
|
@@ -32,8 +30,6 @@ Each data sample contains:
|
|
| 32 |
- `context`: Supporting text passage from which the answer is derived
|
| 33 |
- `answer`: Answer span within the context
|
| 34 |
|
| 35 |
-
---
|
| 36 |
-
|
| 37 |
## ⚙️ Dataset Creation
|
| 38 |
|
| 39 |
**Pipeline:**
|
|
@@ -45,8 +41,6 @@ Each data sample contains:
|
|
| 45 |
|
| 46 |
**Licensing:** All data are collected from public-domain Vietnamese government/service portals and released under CC BY 4.0.
|
| 47 |
|
| 48 |
-
---
|
| 49 |
-
|
| 50 |
## 📊 Quality Evaluation
|
| 51 |
|
| 52 |
A fine-tuned `VietAI/vit5-base` model trained on HVU_QA achieved:
|
|
@@ -60,8 +54,6 @@ A fine-tuned `VietAI/vit5-base` model trained on HVU_QA achieved:
|
|
| 60 |
|
| 61 |
These results confirm that HVU_QA is a high-quality resource for developing robust FAQ-style question generation models.
|
| 62 |
|
| 63 |
-
---
|
| 64 |
-
|
| 65 |
## 📁 Data Fields
|
| 66 |
|
| 67 |
```
|
|
@@ -74,8 +66,6 @@ These results confirm that HVU_QA is a high-quality resource for developing robu
|
|
| 74 |
```
|
| 75 |
> All data files are UTF-8 encoded and ready for use in NLP pipelines.
|
| 76 |
|
| 77 |
-
---
|
| 78 |
-
|
| 79 |
## ⚡ How to Use
|
| 80 |
|
| 81 |
### 📦 Install Dependencies
|
|
@@ -87,7 +77,6 @@ pip install datasets transformers sentencepiece safetensors accelerate evaluate
|
|
| 87 |
*(Install PyTorch separately from [pytorch.org](https://pytorch.org) if not installed yet.)*
|
| 88 |
|
| 89 |
---
|
| 90 |
-
|
| 91 |
### 📥 Load Dataset from Hugging Face Hub
|
| 92 |
|
| 93 |
```python
|
|
@@ -96,9 +85,7 @@ from datasets import load_dataset
|
|
| 96 |
ds = load_dataset("DANGDOCAO/GeneratingQuestions", split="train")
|
| 97 |
print(ds[0])
|
| 98 |
```
|
| 99 |
-
|
| 100 |
---
|
| 101 |
-
|
| 102 |
## 🚀 Example Usage
|
| 103 |
|
| 104 |
### 🔹 Fine-tuning
|
|
@@ -116,7 +103,6 @@ This will:
|
|
| 116 |
*(Or download the pre-trained model: [t5-viet-qg-finetuned](https://huggingface.co/datasets/DANGDOCAO/GeneratingQuestions/tree/main).)*
|
| 117 |
|
| 118 |
---
|
| 119 |
-
|
| 120 |
### 🔹 Generating Questions
|
| 121 |
|
| 122 |
```bash
|
|
@@ -144,7 +130,6 @@ Number of questions: 5
|
|
| 144 |
- `top_k`, `top_p`, `temperature`, `no_repeat_ngram_size`, `repetition_penalty`
|
| 145 |
|
| 146 |
---
|
| 147 |
-
|
| 148 |
## 📌 Citation
|
| 149 |
|
| 150 |
If you use **HVU_QA** in your research, please cite:
|
|
|
|
| 17 |
|
| 18 |
**HVU_QA** is an open-source Vietnamese Question–Context–Answer (QCA) corpus and supporting tools for building FAQ-style question generation systems in low-resource languages. The dataset was created using a fully automated pipeline that combines web crawling from trustworthy sources, semantic tag-based extraction, and AI-assisted filtering to ensure high factual accuracy.
|
| 19 |
|
|
|
|
|
|
|
| 20 |
## 📋 Dataset Description
|
| 21 |
|
| 22 |
- **Language:** Vietnamese
|
|
|
|
| 30 |
- `context`: Supporting text passage from which the answer is derived
|
| 31 |
- `answer`: Answer span within the context
|
| 32 |
|
|
|
|
|
|
|
| 33 |
## ⚙️ Dataset Creation
|
| 34 |
|
| 35 |
**Pipeline:**
|
|
|
|
| 41 |
|
| 42 |
**Licensing:** All data are collected from public-domain Vietnamese government/service portals and released under CC BY 4.0.
|
| 43 |
|
|
|
|
|
|
|
| 44 |
## 📊 Quality Evaluation
|
| 45 |
|
| 46 |
A fine-tuned `VietAI/vit5-base` model trained on HVU_QA achieved:
|
|
|
|
| 54 |
|
| 55 |
These results confirm that HVU_QA is a high-quality resource for developing robust FAQ-style question generation models.
|
| 56 |
|
|
|
|
|
|
|
| 57 |
## 📁 Data Fields
|
| 58 |
|
| 59 |
```
|
|
|
|
| 66 |
```
|
| 67 |
> All data files are UTF-8 encoded and ready for use in NLP pipelines.
|
| 68 |
|
|
|
|
|
|
|
| 69 |
## ⚡ How to Use
|
| 70 |
|
| 71 |
### 📦 Install Dependencies
|
|
|
|
| 77 |
*(Install PyTorch separately from [pytorch.org](https://pytorch.org) if not installed yet.)*
|
| 78 |
|
| 79 |
---
|
|
|
|
| 80 |
### 📥 Load Dataset from Hugging Face Hub
|
| 81 |
|
| 82 |
```python
|
|
|
|
| 85 |
ds = load_dataset("DANGDOCAO/GeneratingQuestions", split="train")
|
| 86 |
print(ds[0])
|
| 87 |
```
|
|
|
|
| 88 |
---
|
|
|
|
| 89 |
## 🚀 Example Usage
|
| 90 |
|
| 91 |
### 🔹 Fine-tuning
|
|
|
|
| 103 |
*(Or download the pre-trained model: [t5-viet-qg-finetuned](https://huggingface.co/datasets/DANGDOCAO/GeneratingQuestions/tree/main).)*
|
| 104 |
|
| 105 |
---
|
|
|
|
| 106 |
### 🔹 Generating Questions
|
| 107 |
|
| 108 |
```bash
|
|
|
|
| 130 |
- `top_k`, `top_p`, `temperature`, `no_repeat_ngram_size`, `repetition_penalty`
|
| 131 |
|
| 132 |
---
|
|
|
|
| 133 |
## 📌 Citation
|
| 134 |
|
| 135 |
If you use **HVU_QA** in your research, please cite:
|