MarcGrumpyOlejak
commited on
Commit
•
4bbd25f
1
Parent(s):
5381e9f
Update README.md
Browse filesAdded name and explanations
README.md
CHANGED
@@ -11,9 +11,12 @@ tags:
|
|
11 |
- merge
|
12 |
|
13 |
---
|
14 |
-
#
|
15 |
|
16 |
-
This is
|
|
|
|
|
|
|
17 |
|
18 |
## Merge Details
|
19 |
### Merge Method
|
@@ -28,6 +31,14 @@ The following models were included in the merge:
|
|
28 |
* [OpenPipe/mistral-ft-optimized-1227](https://huggingface.co/OpenPipe/mistral-ft-optimized-1227)
|
29 |
* [VAGOsolutions/SauerkrautLM-7b-LaserChat](https://huggingface.co/VAGOsolutions/SauerkrautLM-7b-LaserChat)
|
30 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
31 |
### Configuration
|
32 |
|
33 |
The following YAML configuration was used to produce this model:
|
|
|
11 |
- merge
|
12 |
|
13 |
---
|
14 |
+
# VerwaltungsAnthologie_clear_simbad_7B
|
15 |
|
16 |
+
This model is used as an intermediate model for future merges.
|
17 |
+
This is a merge of 4 pre-trained language models based upon Mistral-7B-v0.1 created using [mergekit](https://github.com/cg123/mergekit).
|
18 |
+
|
19 |
+
In combination with [DiscoLM_German_7b_v1](https://huggingface.co/DiscoResearch/DiscoLM_German_7b_v1) this 'clear'-model is the 'base' model to build the successor of my first 'VA_Disco_7B': [VerwaltungsAnthologie_Disco_7B](https://huggingface.co/MarcGrumpyOlejak/VerwaltungsAnthologie_Disco_7B)
|
20 |
|
21 |
## Merge Details
|
22 |
### Merge Method
|
|
|
31 |
* [OpenPipe/mistral-ft-optimized-1227](https://huggingface.co/OpenPipe/mistral-ft-optimized-1227)
|
32 |
* [VAGOsolutions/SauerkrautLM-7b-LaserChat](https://huggingface.co/VAGOsolutions/SauerkrautLM-7b-LaserChat)
|
33 |
|
34 |
+
### Explanations of used models
|
35 |
+
|
36 |
+
* Main focus of this "clear" model is the [hiig-piai/simba-01d](https://huggingface.co/hiig-piai/simba-01d-ftb) (strangley without a 'v') by the Humboldt Institute for Internet and Society has built this model for "simplified language" (Leichte Sprache in german).
|
37 |
+
* The [DRXD1000/Phoenix](https://huggingface.co/DRXD1000/Phoenix) got finetuned with many german texts of law – it can even "hallucinate" almost perfect URL of the official archive of german laws: [Gesetze im Internet]](https://www.gesetze-im-internet.de/)
|
38 |
+
* [OpenPipe/mistral-ft-optimized-1227](https://huggingface.co/OpenPipe/mistral-ft-optimized-1227) performed best using mixed languages in combination with [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) as base model.
|
39 |
+
* [VAGOsolutions/SauerkrautLM-7b-LaserChat](https://huggingface.co/VAGOsolutions/SauerkrautLM-7b-LaserChat) has a wider range of colloquial german language.
|
40 |
+
* [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) is the base model – funny but true – only using [OpenPipe/mistral-ft-optimized-1227](https://huggingface.co/OpenPipe/mistral-ft-optimized-1227) as base model is not as good as combining both.
|
41 |
+
|
42 |
### Configuration
|
43 |
|
44 |
The following YAML configuration was used to produce this model:
|