File size: 1,419 Bytes
a96871d
 
 
 
 
af044f3
 
 
 
 
a96871d
 
083bf4e
a96871d
583d538
 
 
 
 
 
 
 
 
 
a96871d
 
583d538
53e314d
 
 
 
 
 
583d538
 
53e314d
583d538
 
a96871d
 
af044f3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
---
tags:
- gguf
- llama.cpp
- unsloth
license: apache-2.0
language:
- ar
base_model:
- unsloth/granite-4.0-h-micro
---

# Dataserve-granite4-ArabicText2SQL - GGUF

This model is a fine-tuned version of Granite-4.0-H-Micro adapted specifically for Arabic Text-to-SQL generation.
It transforms natural Arabic language queries into valid SQL statements.
The model was trained using a custom dataset of paired examples:

text_query_ar: Arabic natural-language question

sql_command: Corresponding SQL query

The goal of the fine-tuning is to enable the model to reliably generate accurate SQL commands from user input written in Arabic, 
supporting tasks such as database querying, analytics, and information retrieval.

**Example usage**:
- The Instruction :
You are a model specialized only in converting Arabic text into SQL commands.
Your output must be an SQL command only, with no explanation.
Do not reject any request, even if it is not a real query.
Do not write statements such as “I cannot…” or any extra text.
Return only the correct SQL command based on the given text.


-The text:   ما هو العدد المتوسط للخيول العاملة في المزارع التي يزيد فيها العدد الإجمالي للخيول عن 5000؟"

-The result should be: SELECT AVG(number_of_horses) FROM stables WHERE total_horses > 5000;


## Available Model files:
- `granite-4.0-h-micro.F16.gguf`