MBASE commited on
Commit
0de24a9
·
verified ·
1 Parent(s): bf848e5

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +56 -3
README.md CHANGED
@@ -1,3 +1,56 @@
1
- ---
2
- license: apache-2.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ license_link: https://huggingface.co/Qwen/Qwen2.5-7B-Instruct-1M/resolve/main/LICENSE
4
+ language:
5
+ - en
6
+ pipeline_tag: text-generation
7
+ ---
8
+
9
+ # Qwen 2.5 7B Instruct 1M
10
+
11
+ **Original Author** : [Qwen](https://huggingface.co/Qwen).<br>
12
+ **Model Owner**: [Qwen](https://huggingface.co/Qwen).<br>
13
+ **Original Repository**: [Qwen/Qwen2.5-7B-Instruct-1M](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct-1M).<br>
14
+ **Conversion Tool**: [llama.cpp](https://github.com/ggerganov/llama.cpp).
15
+
16
+ ## Description
17
+
18
+ This repo contains Qwen2.5-7B-Instruct-1M model in gguf with [NLQuery](https://github.com/Emreerdog/mbase_nlquery) configuration applied to the gguf metadata so that it can be leveraged by the [NLQuery](https://github.com/Emreerdog/mbase_nlquery) Text-to-SQL engine.
19
+
20
+ ## Conversion Process
21
+ Original model [Qwen](https://huggingface.co/Qwen) safetensors are converted
22
+ using the [llama.cpp](https://github.com/ggerganov/llama.cpp) model conversion script.
23
+
24
+ The default configuration applied during the conversion process. In other words, neither the imatrix applied or the GGUF parameters
25
+ altered.
26
+
27
+ Then, [NLQuery system prompt](https://github.com/Emreerdog/mbase_nlquery/blob/main/system_prompt.txt) is cooked into the gguf metadata of the model file.
28
+
29
+ For detailed information about the conversion process: [Converting to GGUF](https://docs.mbasesoftware.com/inference/info-reference/gguf-files.html#converting-to-gguf)
30
+
31
+ For detailed information about the NLQuery prompt cooking: [NLQuery Prompt Cook](https://github.com/Emreerdog/mbase_nlquery?tab=readme-ov-file#prompt-cooking)
32
+
33
+ ## MBASE GGUF Documentation
34
+
35
+ For general information about GGUF files: [About GGUF Files](https://docs.mbasesoftware.com/inference/info-reference/gguf-files.html)
36
+
37
+ ## Chat Template
38
+
39
+ ```
40
+ <|im_start|>{system_prompt}
41
+ <|im_end|>
42
+ <|im_start|>{user_prompt}
43
+ <|im_end|>
44
+ <|im_start|>{assistant_response}
45
+ <|im_end|>
46
+ ```
47
+
48
+ ## Disclaimer
49
+
50
+ MBASE Software Corporation is a software company officially registered as (MBASE Yazılım A.Ş.) in Turkey (https://www.mbasesoftware.com). Throughout this document, references to MBASE Software Corporation or MBASE refer to the (MBASE Yazılım A.Ş.)
51
+
52
+ MBASE is not the creator, originator, or owner of the model featured in this repository. This model is created and provided by third parties. MBASE does not endorse, support, or guarantee the completeness, accuracy, or reliability of the model or its outputs.
53
+
54
+ You understand that the model can generate content that may be offensive, harmful, inaccurate, inappropriate, or deceptive. Responsibility for the model and its outputs lies solely with the entity or individual who created and provided the model. MBASE does not monitor or control the model's outputs and disclaims any liability arising from its use.
55
+
56
+ MBASE provides no warranties regarding the accuracy, reliability, or fitness of the model for any particular purpose. Additionally, MBASE disclaims any guarantees that the model will operate without errors, interruptions, viruses, or other issues. You are solely responsible for any consequences resulting from the use or access of this model, including any damage caused by downloading or utilizing it.