---
license: apache-2.0
datasets:
- b-mc2/sql-create-context
language:
- en
library_name: transformers
---
# Generate SQL from text - Squeal


Please use the code below as an example for how to use this model.



```python
import torch
from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig

def load_model(model_name):
    # Load tokenizer and model with QLoRA configuration
    compute_dtype = getattr(torch, 'float16')

    bnb_config = BitsAndBytesConfig(
        load_in_4bit=True,
        bnb_4bit_quant_type='nf4',
        bnb_4bit_compute_dtype=compute_dtype,
        bnb_4bit_use_double_quant=False,
    )

    model = AutoModelForCausalLM.from_pretrained(
        model_name,
        device_map={"": 0},
        quantization_config=bnb_config
    )


    # Load Tokenizer
    tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
    tokenizer.pad_token = tokenizer.eos_token
    tokenizer.padding_side = "right"

    return model, tokenizer

model, tokenizer = load_model('vagmi/squeal')

prompt = "<s>[INST] Output SQL for the given table structure \n \
  CREATE TABLE votes (contestant_number VARCHAR, num_votes int); \
  CREATE TABLE contestants (contestant_number VARCHAR, contestant_name VARCHAR); \
  What is the contestant number and name of the contestant who got least votes?[/INST]"
pipe = pipeline(task="text-generation", 
                model=model, 
                tokenizer=tokenizer, 
                max_length=200,
                device_map='auto', )
result = pipe(prompt)
print(result[0]['generated_text'][len(prompt):-1])
```

## How I built it?

Watch me build this model.

https://www.youtube.com/watch?v=PNFhAfxR_d8

Here is the notebook I used to train this model.

https://colab.research.google.com/drive/1jYX8AlRMTY7F_dH3hCFM4ljg5qEmCoUe#scrollTo=IUILKaGWhBxS