Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/run-llama/modal_finetune_sql
https://github.com/run-llama/modal_finetune_sql
Last synced: 9 days ago
JSON representation
- Host: GitHub
- URL: https://github.com/run-llama/modal_finetune_sql
- Owner: run-llama
- License: mit
- Created: 2023-08-16T07:19:55.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2023-11-22T15:59:43.000Z (12 months ago)
- Last Synced: 2024-10-18T21:17:01.536Z (16 days ago)
- Language: Jupyter Notebook
- Size: 115 KB
- Stars: 295
- Watchers: 5
- Forks: 48
- Open Issues: 4
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-llama-resources - Finetuning LLaMa + Text-to-SQL - tune LLaMa 2 7B on a Text-to-SQL dataset (For specific usage Model/ Finetuned model)
- Awesome-Text2SQL - modal_finetune_sql
README
# Finetuning LLaMa + Text-to-SQL
This walkthrough shows you how to fine-tune LLaMa 2 7B on a Text-to-SQL dataset, and then use it for inference against
any database of structured data using LlamaIndex.
Check out our full blog here: https://medium.com/llamaindex-blog/easily-finetune-llama-2-for-your-text-to-sql-applications-ecd53640e10d
This code is taken and adapted from the Modal `doppel-bot` repo: https://github.com/modal-labs/doppel-bot.
### Stack
- LlamaIndex
- Modal
- Hugging Face datasets
- OpenLLaMa
- Peft
### Setup
To get started, clone or fork this repo:
```bash
git clone https://github.com/run-llama/modal_finetune_sql.git
```
### Steps for Running
Please load the notebook `tutorial.ipynb` for full instructions.
```bash
cd modal_finetune_sql
jupyter notebook tutorial.ipynb
```
In the meantime you can run each step individually as below:
Loading data:
`modal run src.load_data_sql`
Finetuning:
`modal run --detach src.finetune_sql`
Inference:
`modal run src.inference_sql_llamaindex::main --query "Which city has the highest population?" --sqlite-file-path "nbs/cities.db"`
(Optional) Downloading model weights:
`modal run src.download_weights --output-dir out_model`