Update README.md
Browse files
README.md
CHANGED
|
@@ -12,6 +12,15 @@ The model is trained on 12 million monolingual ancient Greek tokens with Masked
|
|
| 12 |
| GRC-LAT | 8.200 | [Digital Fragmenta Historicorum Graecorum project](https://www.dfhg-project.org/) |
|
| 13 |
| GRC-KAT <br>GRC-ENG <br>GRC-LAT<br>GRC-ITA<br>GRC-POR | 4.000 | [UGARIT Translation Alignment Editor](https://ugarit.ialigner.com/ ) |
|
| 14 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
If you use this model, please cite our paper:
|
| 16 |
<pre>
|
| 17 |
@misc{yousef_palladino_wright_berti_2022,
|
|
|
|
| 12 |
| GRC-LAT | 8.200 | [Digital Fragmenta Historicorum Graecorum project](https://www.dfhg-project.org/) |
|
| 13 |
| GRC-KAT <br>GRC-ENG <br>GRC-LAT<br>GRC-ITA<br>GRC-POR | 4.000 | [UGARIT Translation Alignment Editor](https://ugarit.ialigner.com/ ) |
|
| 14 |
|
| 15 |
+
### Model Performance
|
| 16 |
+
| Languages | Alignment Error Rate |
|
| 17 |
+
|:---------:|:--------------------:|
|
| 18 |
+
| GRC-ENG | 19.73% (IterMax) |
|
| 19 |
+
| GRC-LAT | 23.91% (IterMax) |
|
| 20 |
+
| GRC-POR | 10.60% (ArgMax) |
|
| 21 |
+
|
| 22 |
+
The gold standard datasets are available on [Github](https://github.com/UgaritAlignment/Alignment-Gold-Standards).
|
| 23 |
+
|
| 24 |
If you use this model, please cite our paper:
|
| 25 |
<pre>
|
| 26 |
@misc{yousef_palladino_wright_berti_2022,
|