gordicaleksa
commited on
Commit
•
e055cbd
1
Parent(s):
9e918ce
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,47 @@
|
|
1 |
-
---
|
2 |
-
license: apache-2.0
|
3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
---
|
4 |
+
|
5 |
+
This repo contains SlovenianGPT - the best open-source base 7B LLM for Slovenian language developed by Aleksa Gordić (see [LinkedIn announcement](https://www.linkedin.com/posts/aleksagordic_sloveniangpt-7b-done-training-after-7-days-activity-7171819162397204480-8U4A?utm_source=share)).
|
6 |
+
|
7 |
+
If you're interested in more powerful instruct models for Slovenian languages feel free to reach out via email (surname.name at gmail com)/[LinkedIn](https://www.linkedin.com/in/aleksagordic).
|
8 |
+
|
9 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/6169e6f39e1ad4ff8735710b/0vTXni9ZIGU0ltC4MLzBu.png" alt="SlovenianGPT" width="450"/>
|
10 |
+
|
11 |
+
SlovenianGPT eval results compared to Mistral 7B, LLaMA 2 7b, and Gemma (also see this [LinkedIn post](https://www.linkedin.com/posts/aleksagordic_happy-to-announce-slovenian-llm-eval-and-activity-7173332520757772288-RB0c?utm_source=share) for more info):
|
12 |
+
|
13 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/6169e6f39e1ad4ff8735710b/bP3DS7-Zr0CR7oTBAWLn1.png" alt="SlovenianGPT eval results" width="750"/>
|
14 |
+
|
15 |
+
Instruct-SlovenianGPT eval results ([LinkedIn post](https://www.linkedin.com/posts/aleksagordic_very-happy-to-announce-sloveniangpt-instruct-activity-7181638113197248512-5jvs?utm_source=share)):
|
16 |
+
|
17 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/6169e6f39e1ad4ff8735710b/EHEh71HvT7eAMMtzROrjm.png" alt="SlovenianGPT eval results" width="750"/>
|
18 |
+
|
19 |
+
Eval was computed using https://github.com/gordicaleksa/slovenian-llm-eval
|
20 |
+
|
21 |
+
The model was trained on tens of billions of Slovenian language tokens and is based off of [Mistral 7B](https://huggingface.co/mistralai/Mistral-7B-v0.1).
|
22 |
+
|
23 |
+
## Notes
|
24 |
+
|
25 |
+
1) SlovenianGPT is a base model and therefore does not have any moderation mechanisms.
|
26 |
+
|
27 |
+
2) Since it's a base model it won't follow your instructions as it's just a powerful autocomplete engine.
|
28 |
+
|
29 |
+
3) If you want an access to much more powerful Slovenian LLMs - feel free to reach out to me via email (surname.name at gmail com)/[LinkedIn](https://www.linkedin.com/in/aleksagordic).
|
30 |
+
|
31 |
+
# Credits
|
32 |
+
|
33 |
+
The data for the project was obtained with the help of [Nikola Ljubešić](https://nljubesi.github.io/).
|
34 |
+
|
35 |
+
**Also a big thank you to the following individuals:**
|
36 |
+
- [**Aleksander Segedi**](https://www.linkedin.com/in/aleksander-segedi-08430936/) - for help around bookkeeping!
|
37 |
+
|
38 |
+
## Citation
|
39 |
+
|
40 |
+
```
|
41 |
+
@article{SlovenianGPT,
|
42 |
+
author = "Gordić Aleksa",
|
43 |
+
title = "SlovenianGPT - an open-source LLM for Slovenian language",
|
44 |
+
year = "2024"
|
45 |
+
howpublished = {\url{https://huggingface.co/gordicaleksa/SlovenianGPT}},
|
46 |
+
}
|
47 |
+
```
|