README.md

---
base_model: t5-small
library_name: peft
---

# Model Card for Model ID

Convert natural language queries into SQL.



## Model Details

### Model Description

A text-to-sql LoRA adapter for m5-small finetuned on the b-mc2/sql-create-context, Clinton/Text-to-sql-v1, and knowrohit07/know_sql datasets.



- **Model type:** text-to-sql
- **Language(s) (NLP):** Python (PyTorch)
- **Finetuned from model:** m5-small

### Model Sources [optional]

<!-- Provide the basic links for the model. -->

- **Repository:** [github: Text2sql-LoRA-T5-small](https://github.com/KookyGhost/Text2sql-LoRA-T5-small.git)

## Training Details

### Training Data

[b-mc2/sql-create-context](https://huggingface.co/datasets/b-mc2/sql-create-context)

[Clinton/Text-to-sql-v1](https://huggingface.co/datasets/Clinton/Text-to-sql-v1)

[knowrohit07/know_sql](https://huggingface.co/datasets/knowrohit07/know_sql)

Train-Eval-Test split: 8:1:1

#### Metrics

<!-- These are the evaluation metrics being used, ideally with a description of why. -->

[ROUGE](https://huggingface.co/metrics/rouge)

### Results

'rouge1': 0.9281907395888822, 'rouge2': 0.8917983416395394, 'rougeL': 0.9182615264948171, 'rougeLsum': 0.9191261027066355

#### Summary

This adapter helps m5-small achieve a ROUGE score of over 90% on text to SQL tasks.


## Environmental Impact

<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->

Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).

- **Hardware Type:** RTX 4000
- **Hours used:** 2
- **Cloud Provider:** Paperspace


- PEFT 0.6.2