Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,43 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
license: mit
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
language:
|
| 3 |
+
- vi
|
| 4 |
+
pipeline_tag: text-classification
|
| 5 |
+
tags:
|
| 6 |
+
- clickbait
|
| 7 |
+
- vietnamese
|
| 8 |
+
- phobert
|
| 9 |
+
- text-classification
|
| 10 |
license: mit
|
| 11 |
---
|
| 12 |
+
|
| 13 |
+
# Vietnamese Clickbait Detection
|
| 14 |
+
|
| 15 |
+
Mô hình phát hiện tiêu đề clickbait tiếng Việt.
|
| 16 |
+
|
| 17 |
+
## Architecture
|
| 18 |
+
|
| 19 |
+
- Encoder: PhoBERT-base
|
| 20 |
+
- Additional layers: CNN and pooling
|
| 21 |
+
- Task: Binary classification
|
| 22 |
+
- Labels:
|
| 23 |
+
- 0: Non-clickbait
|
| 24 |
+
- 1: Clickbait
|
| 25 |
+
|
| 26 |
+
## Inputs
|
| 27 |
+
|
| 28 |
+
The model receives:
|
| 29 |
+
|
| 30 |
+
- News title
|
| 31 |
+
- Lead paragraph or article context
|
| 32 |
+
|
| 33 |
+
## Limitations
|
| 34 |
+
|
| 35 |
+
The model was trained on Vietnamese news data and may not perform reliably on:
|
| 36 |
+
|
| 37 |
+
- Other languages
|
| 38 |
+
- Very short or incomplete inputs
|
| 39 |
+
- Domains that differ significantly from the training dataset
|
| 40 |
+
|
| 41 |
+
## Intended use
|
| 42 |
+
|
| 43 |
+
This repository stores the trained checkpoint used by the Streamlit demonstration application.
|