A framework for AI-assisted program synthesis. Given a problem description and some input-output examples, the framework generates a program that solves the problem.
You can find an in-depth discussion of this tool, the philosophy it implements and its usage in our paper, Fully Autonomous Programming with Large Language Models. Consider citing it if you use SEIDR in your research.
from seidr import dev
help(dev)
The experiments are contained in benchmark.py
and benchmark_humaneval.py
files. When you run this file, the AI-generated programs are commited to a dedicated github repository, while the metrics (i.e. how many tests every program passes) will be logged in your Weights and Biases
- Create an account on Weights and Biases
- Install the Weights and Biases library
- Run
wandb login
and follow the instructions
- Go to github, log in to the account that's going to push AI-generated code. Remember the $username and $email for that account.
- Go here and generate an access $token
- Set
GIT_USER
to "Bot" or whatever the name of the committer shall be - Set
GIT_EMAIL
to $email - Set
GIT_REMOTE
to https://$username:$[email protected]/$repo
Note that you can use a non-GitHub git hosting.
OpenAI account is needed with access to gpt-3.5-turbo
and
an OPENAI_API_KEY
environment variable
set to your OpenAI API access token.
Run Ollama with CodeLlama or another model locally
or on a server.
In the latter case, start the Ollama server with the following commands and note the URL:PORT
pair:
OLLAMA_HOST=URL:PORT ollama serve &
OLLAMA_HOST=URL:PORT ollama pull codellama:34b-instruct &
Example .config
file layout:
# Github
export GIT_REMOTE=https://USERNAME:[email protected]/SOLUTIONS_REPO
export GIT_USER=...
export GIT_EMAIL=...
# Data
export DATA_PATH=...
# OpenAI
export OPENAI_API_KEY=...
export OPENAI_ORG=...
# WandB
export WANDB_ENTITY=...
export WANDB_DIR=...
If you're using Slurm, write a run.sh
file with python benchmark.py
and run it with sbatch run.sh --array=1-500
.
If not, run TASK_ID=n python benchmark.py
to re-run one of our experiments exactly, or set the parameters yourself:
For example, for basement problem in PSB2, run SEIDR without lexicase selection as follows:
python3 benchmark.py \
--task_id 202 \
--problem basement \
--language C++ \
--max_programs 100 \
--drafts_per_prompt 2 \
--explanations_per_program 2 \
--repairs_per_explanation 2 \
--beam_width 2 \
--log INFO \
--lexicase_selection False \
--dataset psb2 \
--model_name gpt-3.5-turbo
To run an example with SEIDR with CodeLlama served by Ollama at URL:PORT
, run the following:
python3 benchmark.py \
--task_id 2202 \
--problem basement \
--language C++ \
--max_programs 100 \
--drafts_per_prompt 2 \
--explanations_per_program 2 \
--repairs_per_explanation 2 \
--beam_width 2 \
--log INFO \
--lexicase_selection False \
--dataset psb2 \
--model_name codellama:34b-instruct \
--ollama_url "http://URL:PORT"
Example Slurm scripts are stored in example_scripts/
and tables with hyperparameters in /config