Update README.md
Browse files
README.md
CHANGED
@@ -24,14 +24,13 @@ SauerkrautLM-v1 is here to empower your language-related endeavors like never be
|
|
24 |
| Model | HF | GPTQ | GGUF | AWQ |
|
25 |
|-------|-------|-------|-------|-------|
|
26 |
| SauerkrautLM-7b-HerO | [Link](https://huggingface.co/VAGOsolutions/SauerkrautLM-7b-HerO) | coming soon | coming soon | coming soon |
|
27 |
-
| SauerkrautLM-7b-HerO-multilingual | [Link](https://huggingface.co/VAGOsolutions/SauerkrautLM-7b-HerO-multilingual) | coming soon | coming soon | coming soon |
|
28 |
|
29 |
## Model Details
|
30 |
**SauerkrautLM-7b-HerO**
|
31 |
|
32 |
**Training Dataset:**
|
33 |
|
34 |
-
SauerkrautLM-7b-HerO
|
35 |
We found, that only a simple translation of training data can lead to unnatural German phrasings.
|
36 |
Data augmentation techniques were used to grant grammatical, syntactical correctness and a more natural German wording in our training data.
|
37 |
|
@@ -149,13 +148,21 @@ neural-chat-7b-v3-1 6.812500
|
|
149 |
|
150 |
|
151 |
**Language Model evaluation Harness**
|
152 |
-
|
153 |
-
|
154 |
|
|
|
155 |
**BBH**
|
156 |
-
![BBH](images/bbh.PNG "SauerkrautLM-7b-HerO BBH")
|
157 |
-
|
158 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
159 |
## Disclaimer
|
160 |
We must inform users that despite our best efforts in data cleansing, the possibility of some such content slipping through cannot be entirely ruled out.
|
161 |
However, we cannot guarantee consistently appropriate behavior. Therefore, if you encounter any issues or come across inappropriate content, we kindly request that you inform us through the contact information provided.
|
|
|
24 |
| Model | HF | GPTQ | GGUF | AWQ |
|
25 |
|-------|-------|-------|-------|-------|
|
26 |
| SauerkrautLM-7b-HerO | [Link](https://huggingface.co/VAGOsolutions/SauerkrautLM-7b-HerO) | coming soon | coming soon | coming soon |
|
|
|
27 |
|
28 |
## Model Details
|
29 |
**SauerkrautLM-7b-HerO**
|
30 |
|
31 |
**Training Dataset:**
|
32 |
|
33 |
+
SauerkrautLM-7b-HerO was trained with mix of German data augmentation and translated data.
|
34 |
We found, that only a simple translation of training data can lead to unnatural German phrasings.
|
35 |
Data augmentation techniques were used to grant grammatical, syntactical correctness and a more natural German wording in our training data.
|
36 |
|
|
|
148 |
|
149 |
|
150 |
**Language Model evaluation Harness**
|
151 |
+
Compared to Aleph Alpha Luminous Models:
|
152 |
+
![Harness](images/luminouscompare.PNG "SauerkrautLM-7b-HerO Harness") *
|
153 |
|
154 |
+
*performed with newest Language Model Evaluation Harness
|
155 |
**BBH**
|
156 |
+
![BBH](images/bbh.PNG "SauerkrautLM-7b-HerO BBH") *
|
157 |
+
*performed with newest Language Model Evaluation Harness
|
158 |
+
**GPT4ALL**
|
159 |
+
Compared to Aleph Alpha Luminous Models, LeoLM and EM_German:
|
160 |
+
![GPT4ALL diagram](images/gpt4alldiagram.PNG "SauerkrautLM-7b-HerO GPT4ALL Diagram") *
|
161 |
+
|
162 |
+
![GPT4ALL table](images/gpt4alltable.PNG "SauerkrautLM-7b-HerO GPT4ALL Table") *
|
163 |
+
**Additional German Benchmark results**
|
164 |
+
![GermanBenchmarks](images/germanbench.PNG "SauerkrautLM-7b-HerO German Benchmarks") *
|
165 |
+
*performed with newest Language Model Evaluation Harness
|
166 |
## Disclaimer
|
167 |
We must inform users that despite our best efforts in data cleansing, the possibility of some such content slipping through cannot be entirely ruled out.
|
168 |
However, we cannot guarantee consistently appropriate behavior. Therefore, if you encounter any issues or come across inappropriate content, we kindly request that you inform us through the contact information provided.
|