Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cc-by-4.0
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
pipeline_tag: text-classification
|
6 |
+
tags:
|
7 |
+
- distilroberta
|
8 |
+
- topic
|
9 |
+
- news
|
10 |
+
---
|
11 |
+
|
12 |
+
# Fine-tuned distilroberta-base for detecting whether news is from a wire service
|
13 |
+
|
14 |
+
# Model Description
|
15 |
+
|
16 |
+
This model is a finetuned distilroberta-base, for classifying whether news articles are from a wire service.
|
17 |
+
|
18 |
+
# How to Use
|
19 |
+
|
20 |
+
```python
|
21 |
+
from transformers import pipeline
|
22 |
+
classifier = pipeline("text-classification", model="dell-research-harvard/wire-classifier")
|
23 |
+
classifier("Washington (AP) Two men died in a car crash")
|
24 |
+
```
|
25 |
+
|
26 |
+
# Training data
|
27 |
+
|
28 |
+
The model was trained on a hand-labelled sample of data from the [NEWSWIRE dataset](https://huggingface.co/datasets/dell-research-harvard/newswire).
|
29 |
+
|
30 |
+
Split|Size
|
31 |
+
-|-
|
32 |
+
Train|1,459
|
33 |
+
Dev|336
|
34 |
+
Test|448
|
35 |
+
|
36 |
+
# Test set results
|
37 |
+
|
38 |
+
Metric|Result
|
39 |
+
-|-
|
40 |
+
F1|0.96
|
41 |
+
Accuracy|
|
42 |
+
Precision|
|
43 |
+
Recall|
|
44 |
+
|
45 |
+
|
46 |
+
# Citation Information
|
47 |
+
|
48 |
+
You can cite this dataset using
|
49 |
+
|
50 |
+
```
|
51 |
+
@misc{silcock2024newswirelargescalestructureddatabase,
|
52 |
+
title={Newswire: A Large-Scale Structured Database of a Century of Historical News},
|
53 |
+
author={Emily Silcock and Abhishek Arora and Luca D'Amico-Wong and Melissa Dell},
|
54 |
+
year={2024},
|
55 |
+
eprint={2406.09490},
|
56 |
+
archivePrefix={arXiv},
|
57 |
+
primaryClass={cs.CL},
|
58 |
+
url={https://arxiv.org/abs/2406.09490},
|
59 |
+
}
|
60 |
+
```
|
61 |
+
|
62 |
+
# Applications
|
63 |
+
|
64 |
+
We applied this model to a century of historical news articles. You can see all the classifications in the [NEWSWIRE dataset](https://huggingface.co/datasets/dell-research-harvard/newswire).
|
65 |
+
|
66 |
+
|