Update README.md
Browse files
README.md
CHANGED
@@ -7,7 +7,7 @@ widget:
|
|
7 |
license: mit
|
8 |
---
|
9 |
|
10 |
-
# GerPT2
|
11 |
|
12 |
German large and small versions of GPT2:
|
13 |
|
@@ -25,6 +25,7 @@ I evaluated both GerPT2-large and the other German GPT2, [dbmdz/german-gpt2](htt
|
|
25 |
| dbmdz/german-gpt2 | 49.47 | 62.92 |
|
26 |
| GerPT2 | 24.78 | 35.33 |
|
27 |
| GerPT2-large | __16.08__ | __23.26__ |
|
|
|
28 |
|
29 |
See the script `evaluate.py` in the [GerPT2 Github repository](https://github.com/bminixhofer/gerpt2) for the code.
|
30 |
|
@@ -108,7 +109,22 @@ This helps a lot on a trial run I did, although I wasn't able to do a full compa
|
|
108 |
|
109 |
GerPT2-large is licensed under the MIT License.
|
110 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
111 |
## Acknowledgements
|
112 |
|
113 |
Thanks to [Hugging Face](https://huggingface.co) for awesome tools and infrastructure.
|
114 |
-
Huge thanks to [Artus Krohn-Grimberghe](https://twitter.com/artuskg) at [LYTiQ](https://www.lytiq.de/) for making this possible by sponsoring the resources used for training.
|
|
|
7 |
license: mit
|
8 |
---
|
9 |
|
10 |
+
# GerPT2
|
11 |
|
12 |
German large and small versions of GPT2:
|
13 |
|
|
|
25 |
| dbmdz/german-gpt2 | 49.47 | 62.92 |
|
26 |
| GerPT2 | 24.78 | 35.33 |
|
27 |
| GerPT2-large | __16.08__ | __23.26__ |
|
28 |
+
| | | |
|
29 |
|
30 |
See the script `evaluate.py` in the [GerPT2 Github repository](https://github.com/bminixhofer/gerpt2) for the code.
|
31 |
|
|
|
109 |
|
110 |
GerPT2-large is licensed under the MIT License.
|
111 |
|
112 |
+
## Citing
|
113 |
+
|
114 |
+
Please cite GerPT2 as follows:
|
115 |
+
|
116 |
+
```
|
117 |
+
@misc{Minixhofer_GerPT2_German_large_2020,
|
118 |
+
author = {Minixhofer, Benjamin},
|
119 |
+
doi = {10.5281/zenodo.5509984},
|
120 |
+
month = {12},
|
121 |
+
title = {{GerPT2: German large and small versions of GPT2}},
|
122 |
+
url = {https://github.com/bminixhofer/gerpt2},
|
123 |
+
year = {2020}
|
124 |
+
}
|
125 |
+
```
|
126 |
+
|
127 |
## Acknowledgements
|
128 |
|
129 |
Thanks to [Hugging Face](https://huggingface.co) for awesome tools and infrastructure.
|
130 |
+
Huge thanks to [Artus Krohn-Grimberghe](https://twitter.com/artuskg) at [LYTiQ](https://www.lytiq.de/) for making this possible by sponsoring the resources used for training.
|