Update README.md

README.md

@@ -7,53 +7,42 @@ tags:
 - unsloth
 - trl
 - sft
-licence: license
+licence: apache-2.0
 ---
 
-# Model Card for onesql-completions3
+# Introduction
 
-This model is a fine-tuned version of [unsloth/Qwen2.5-Coder-32B-Instruct-bnb-4bit](https://huggingface.co/unsloth/Qwen2.5-Coder-32B-Instruct-bnb-4bit).
-It has been trained using [TRL](https://github.com/huggingface/trl).
+This model specializes on the Text-to-SQL task. It is finetuned from the quantized version of [Qwen2.5-Coder-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct).
+The model has an EX score of 63.33 and R-VES score of 60.02 on the [BIRD leaderboard](https://bird-bench.github.io/).
 
-## Quick start
+# Quick start
 
-```python
-from transformers import pipeline
-
-question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
-generator = pipeline("text-generation", model="onekq-ai/onesql-completions3", device="cuda")
-output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
-print(output["generated_text"])
-```
-
-## Training procedure
-
- 
-
-
-This model was trained with SFT.
+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.
 
-### Framework versions
-
-- TRL: 0.13.0
-- Transformers: 4.47.1
-- Pytorch: 2.5.1
-- Datasets: 3.2.0
-- Tokenizers: 0.21.0
-
-## Citations
-
-
-
-Cite TRL as:
-    
-```bibtex
-@misc{vonwerra2022trl,
-	title        = {{TRL: Transformer Reinforcement Learning}},
-	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
-	year         = 2020,
-	journal      = {GitHub repository},
-	publisher    = {GitHub},
-	howpublished = {\url{https://github.com/huggingface/trl}}
-}
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
+from peft import PeftModel
+
+model_name = "unsloth/Qwen2.5-Coder-32B-Instruct-bnb-4bit"
+adapter_name = "onekq-ai/OneSQL-v0.1-Qwen-32B"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+tokenizer.padding_side = "left"
+model = PeftModel.from_pretrained(AutoModelForCausalLM.from_pretrained(model_name, device_map="auto"), adapter_name).to("cuda")
+
+generator = pipeline("text-generation", model=model, tokenizer=tokenizer, return_full_text=False)
+
+prompt = """
+CREATE TABLE students (
+    id INTEGER PRIMARY KEY,
+    name TEXT,
+    age INTEGER,
+    grade TEXT
+);
+
+-- Find the three youngest students
+SELECT """
+
+result = generator(f"<|im_start|>system\nYou are a SQL expert. Return code only.<|im_end|>\n<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant\n")[0]
+print(result["generated_text"])
 ```