DavidGF commited on
Commit
b2f2b9a
1 Parent(s): 2153efc

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +97 -0
README.md ADDED
@@ -0,0 +1,97 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: llama3.1
3
+ language:
4
+ - de
5
+ - en
6
+ - it
7
+ - fr
8
+ - pt
9
+ - es
10
+ tags:
11
+ - spectrum
12
+ ---
13
+
14
+ ![Llama-3.1-SauerkrautLM-8b-Instruct]( https://vago-solutions.ai/wp-content/uploads/2024/07/Llama3.1-SauerkrautLM.png "Llama-3.1-SauerkrautLM-8b-Instruct")
15
+ ## VAGO solutions Llama-3.1-SauerkrautLM-8b-Instruct-awq
16
+
17
+ **Fine-tuned Model** - *to showcase the potential of resource-efficient Fine-Tuning of Large Language Models using **Spectrum Fine-Tuning***
18
+
19
+ Introducing **Llama-3.1-SauerkrautLM-8b-Instruct-awq ** – our Sauerkraut AWQ version of the powerful [VAGOsolutions/Llama-3.1-SauerkrautLM-8b-Instruct](https://huggingface.co/VAGOsolutions/Llama-3.1-SauerkrautLM-8b-Instruct)!
20
+
21
+ - Fine-tuning on German-English data with [**Spectrum**](https://github.com/cognitivecomputations/spectrum) Fine-Tuning **targeting 25% of the layers.**
22
+ - Utilized unique German-English Sauerkraut Mix v2
23
+ - Implemented bespoke, precision-engineered fine-tuning approach
24
+
25
+ # Table of Contents
26
+ 1. [Overview of all Llama-3.1-SauerkrautLM-8b-Instruct](#all-Llama-3.1-SauerkrautLM-8b-Instruct)
27
+ 2. [Model Details](#model-details)
28
+ - [Training procedure](#training-procedure)
29
+ 3. [Evaluation](#evaluation)
30
+ 5. [Disclaimer](#disclaimer)
31
+ 6. [Contact](#contact)
32
+ 7. [Collaborations](#collaborations)
33
+ 8. [Acknowledgement](#acknowledgement)
34
+
35
+ ## All Llama-3.1-SauerkrautLM-8b-Instruct
36
+
37
+ | Model | HF | EXL2 | GGUF | AWQ |
38
+ |-------|-------|-------|-------|-------|
39
+ | Llama-3.1-SauerkrautLM-8b-Instruct | [Link](https://huggingface.co/VAGOsolutions/Llama-3.1-SauerkrautLM-8b-Instruct) | coming soon | coming soon | [Link](https://huggingface.co/VAGOsolutions/Llama-3.1-SauerkrautLM-8b-Instruct-awq) |
40
+
41
+ ## Model Details
42
+ **Llama-3.1-SauerkrautLM-8b-Instruct**
43
+ - **Model Type:** Llama-3.1-SauerkrautLM-8b-Instruct is a fine-tuned Model based on [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/mistralai/meta-llama/Meta-Llama-3.1-8B-Instruct)
44
+ - **Language(s):** German, English
45
+ - **License:** llama3.1
46
+ - **Contact:** [VAGO solutions](https://vago-solutions.ai)
47
+
48
+ ## Training Procedure
49
+
50
+ This model showcases the potential of resource-efficient fine-tuning of large language models using Spectrum Fine-Tuning. Here's a brief on the procedure:
51
+
52
+ **Fine-tuning on German-English Data**:
53
+
54
+ - Utilized Spectrum Fine-Tuning, targeting 25% of the model's layers
55
+ - Introduced the model to a unique German-English Sauerkraut Mix v2
56
+ - Implemented a bespoke, precision-engineered fine-tuning approach
57
+
58
+ **Sauerkraut Mix v2**:
59
+
60
+ - Premium Dataset for Language Models, focusing on German and English
61
+ - Meticulously selected, high-quality dataset combinations
62
+ - Cutting-edge synthetic datasets created using proprietary, high-precision generation techniques
63
+
64
+ ## Objective and Results
65
+
66
+ The primary goal of this training was to demonstrate that with Spectrum Fine-Tuning targeting 25% of the layers, a 8 billion parameter model can significantly enhance the capabilities while using a fraction of the resources of the classic fine-tuning approach.
67
+
68
+ The model has substantially improved skills in German and English, as demonstrated by impressive benchmarks on the new Hugging Face leaderboard.
69
+
70
+ **Spectrum Fine-Tuning can efficiently enhance a large language model's capabilities in multiple languages while preserving the majority of its previously acquired knowledge.**
71
+
72
+ ## Evaluation
73
+
74
+ **AGIEVAL**
75
+ ![Llama-3.1-SauerkrautLM-8b-Instruct-AGIEVAL]( https://vago-solutions.ai/wp-content/uploads/2024/07/llama3.1-agieval1.png "Llama-3.1-SauerkrautLM-8b-Instruct-AGIEVAL")
76
+
77
+ **GPT4ALL**
78
+ ![Llama-3.1-SauerkrautLM-8b-Instruct-GPT4ALL]( https://vago-solutions.ai/wp-content/uploads/2024/07/llama3.1-GPT4ALL1.png "Llama-3.1-SauerkrautLM-8b-Instruct-GPT4ALL")
79
+
80
+ **TRUTHFULQA**
81
+ ![Llama-3.1-SauerkrautLM-8b-Instruct-TRUTHFULQA]( https://vago-solutions.ai/wp-content/uploads/2024/07/llama3.1-TQA1.png "Llama-3.1-SauerkrautLM-8b-Instruct-TRUTHFULQA")
82
+
83
+ **OPENLEADERBOARD 2**
84
+ ![Llama-3.1-SauerkrautLM-8b-Instruct-OPENLEADERBOARD]( https://vago-solutions.ai/wp-content/uploads/2024/07/llama3.1-HF21.png "Llama-3.1-SauerkrautLM-8b-Instruct-OPENLEADERBOARD")
85
+
86
+
87
+ ## Disclaimer
88
+ We must inform users that despite our best efforts in data cleansing, the possibility of uncensored content slipping through cannot be entirely ruled out. 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. Additionally, it is essential to understand that the licensing of these models does not constitute legal advice. We are not held responsible for the actions of third parties who utilize our models.
89
+
90
+ ## Contact
91
+ If you are interested in customized LLMs for business applications, please get in contact with us via our website. We are also grateful for your feedback and suggestions.
92
+
93
+ ## Collaborations
94
+ We are also keenly seeking support and investment for our startup, VAGO solutions where we continuously advance the development of robust language models designed to address a diverse range of purposes and requirements. If the prospect of collaboratively navigating future challenges excites you, we warmly invite you to reach out to us at [VAGO solutions](https://vago-solutions.ai)
95
+
96
+ ## Acknowledgement
97
+ Many thanks to [meta-llama](https://huggingface.co/meta-llama) for providing such a valuable model to the Open-Source community.