https://github.com/arya920/natural_language_to_sql_queries

The task of this project is to Convert Natural Language to SQL Queries
https://github.com/arya920/natural_language_to_sql_queries

deepseek-coder gemma-2b huggingface-datasets huggingface-transformer lora mysql-database peft-fine-tuning-llm sqllite streamlit transformers

Last synced: 13 days ago
JSON representation

The task of this project is to Convert Natural Language to SQL Queries

• Host: GitHub
• URL: https://github.com/arya920/natural_language_to_sql_queries
• Owner: Arya920
• Created: 2024-07-25T06:44:16.000Z (7 months ago)
• Default Branch: master
• Last Pushed: 2024-08-02T11:42:45.000Z (6 months ago)
• Last Synced: 2024-12-04T11:06:24.281Z (2 months ago)
• Topics: deepseek-coder, gemma-2b, huggingface-datasets, huggingface-transformer, lora, mysql-database, peft-fine-tuning-llm, sqllite, streamlit, transformers
• Language: Python
• Homepage:
• Size: 2.64 MB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
---

title: Streanlit GEMMA 2B & Deepseek Coder 1.3B

emoji: 📉

colorFrom: gray

colorTo: gray

sdk: streamlit

sdk_version: 1.31.1

app_file: app.py

run_code: streamlit run app.py

pinned: false

license: Apache 2.0

---

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

### Speak2SQl 

This web app is built to convert the natural language to sql queries. It is powered by 2 state of the art Transformer model, 

- GEMMA 2B

- Deepseek Coder 1.3B

Both the models are fine tuned over 3000 rows and have been tested againest 100 seperate rows. The evaluation metrics used for evaluating these models are ~ 

- BLEU SCORE

- SCAR BLEU SCORE

- EM RATIO

Here BLEU SCORE considers the ngram probability and checks that if the predicted text matches with the actual text over n-gram, whereas EM Ratio checks that the predicted query exactly matches with the actual query or not.