sqllama-V0 / README.md
matt-tries-dl's picture
final
2a75437
|
raw
history blame
690 Bytes

Setup

This is for the gcp vm. Mine came installed with python3.7 and cuda 11.3

Create a venv

python3 -m venv .venv
source .venv/bin/activate You may have to install python3-venv to the machine.

Install

pip install -r requirements.txt
pip uninstall -r uninstall.txt

I had to uninstall some cuda stuff that torch installed to make this work.

References

https://github.com/tloen/alpaca-lora https://huggingface.co/docs/transformers/main/en/model_doc/llama#llama https://huggingface.co/docs/transformers/index https://github.com/salesforce/WikiSQL https://github.com/huggingface/peft

https://arxiv.org/pdf/1910.13461.pdf https://arxiv.org/pdf/1808.06226.pdf