onekq-ai/OneSQL-v0.1-Qwen-3B-GGUF

Introduction

This model is the GGUF version of OneSQL-v0.1-Qwen-3B. You can also find it on Ollama.

Performances

The self-evaluation EX score of the original model is 43.35 (compared to 63.33 by the 32B model on the BIRD leaderboard. Below is the self-evaluation results for each quantization.

Quantization	EX score
Q4_0	16.83
Q4_1	21.85
Q4_K_S	22.49
Q4_K_M	21.85
Q5_0	23.40
Q5_1	23.53
Q5_K_S	22.77
Q5_K_M	23.73
Q6_K	24.51
Q8_0	24.90

Quick start

To use this model, craft your prompt to start with your database schema in the form of CREATE TABLE, followed by your natural language query preceded by --. Make sure your prompt ends with SELECT in order for the model to finish the query for you. There is no need to set other parameters like temperature or max token limit.

PROMPT="CREATE TABLE students (
    id INTEGER PRIMARY KEY,
    name TEXT,
    age INTEGER,
    grade TEXT
);

-- Find the three youngest students
SELECT "

ollama run onekq-ai/OneSQL-v0.1-Qwen:3B-Q8_0 "$PROMPT"

The model response is the finished SQL query without SELECT

* FROM students ORDER BY age ASC LIMIT 3

Caveats

• The performance drop from the original model is due to quantization itself, and the lack of beam search support in llama.cpp framework. Use at your own discretion.
• The 2-bit and 3-bit quantizations suffer from repetitive and unrelevant output token, hence are not recommended for usage.