Skip to content

Code for SIGMOD DEEM'24 paper titled: Towards Efficient Data Wrangling with LLMs using Code Generation.

Notifications You must be signed in to change notification settings

effyli/efficient_llm_data_wrangling

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 

Repository files navigation

efficient_llm_data_wrangling

This github repo contains code for our SIGMOD DEEM'24 paper: Towards Efficient Data Wrangling with LLMs using Code Generation.

Following work by Narayan et al., we use the same set of benchmark datasets. You can clone the repo and download the data by using:

git clone [email protected]:effyli/efficient_llm_data_wrangling.git
mkdir data/
wget https://fm-data-tasks.s3.us-west-1.amazonaws.com/datasets.tar.gz -P data
tar xvf data/datasets.tar.gz -C data/

To run the script, first setup the data_dir environmental variable by using:

export DATASET_PATH="$PWD/data/datasets"

One example command of calling the script:

python run_wrangler.py \
  --data_dir data/datasets/data_transformation/benchmark-bing-query-logs-semantics  \
  --num_trials 3  \
  --seed 42 \
  --k 3 \
  --d 0 \
  --num_iter 1

About

Code for SIGMOD DEEM'24 paper titled: Towards Efficient Data Wrangling with LLMs using Code Generation.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages