Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
datasets:
|
3 |
+
- oscar
|
4 |
+
language:
|
5 |
+
- da
|
6 |
+
widget:
|
7 |
+
- text: Der var engang
|
8 |
+
---
|
9 |
+
|
10 |
+
# What is this?
|
11 |
+
|
12 |
+
GPT-2 model (small version, 124 M parameters) for Danish text generation.
|
13 |
+
|
14 |
+
# Model training
|
15 |
+
|
16 |
+
The model is trained using the Danish part of the [oscar dataset](https://huggingface.co/datasets/oscar) ('unshuffled_deduplicated_da') and a context length of 1024 tokens.
|
17 |
+
|
18 |
+
The model is initilized from the English [GPT-2 small model](https://huggingface.co/gpt2) with new word token embeddings created for Danish using [WECHSEL](https://github.com/CPJKU/wechsel).
|
19 |
+
|
20 |
+
Initially, only the word token embeddings are trained using 50.000 samples. Finally, the whole model is trained using 1.000.000 samples.
|
21 |
+
|
22 |
+
Model training is carried out on a 8 GB GPU.
|
23 |
+
|
24 |
+
# Notes
|
25 |
+
|
26 |
+
This is a pre-trained model, for optimal performance it should be finetuned for new tasks.
|