Qwen2.5-0.5B-SQL LoRA Adapter

This model is a LoRA (Low-Rank Adaptation) adapter for Qwen2.5-0.5B-Instruct, specifically fine-tuned to generate SQL queries from natural language questions and database schemas.

Model Details

• Base Model: Qwen/Qwen2.5-0.5B-Instruct
• Task: Text-to-SQL
• Training Data: b-mc2/sql-create-context
• Language: English

Quick Start (How to use)

To use this adapter, you need to load the base model first and then apply the LoRA weights.

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel

model_id = "Qwen/Qwen2.5-0.5B-Instruct"
adapter_id = "azeemazam/Qwen2.5-0.5B-SQL"

tokenizer = AutoTokenizer.from_pretrained(adapter_id)
base_model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.float16, device_map='auto')
model = PeftModel.from_pretrained(base_model, adapter_id)

def generate_sql(schema, question):
    messages = [
        {"role": "user", "content": f"Generate SQL.\\n\\nDatabase Schema:\\n{schema}\\n\\nQuestion:\\n{question}"}    
    ]
    prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
    inputs = tokenizer(prompt, return_tensors='pt').to(model.device)
    
    outputs = model.generate(**inputs, max_new_tokens=150)
    return tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True)

schema = "CREATE TABLE employees (id INT, name TEXT, salary INT)"
question = "Who earns more than 50000?"
print(generate_sql(schema, question))