config.yaml

iteration_num: 2
batch_size: 2
data_parallel_size: 32
devoid_duplication: false
llm_model: "gpt-3.5-turbo"
timeout: 10
data_dir: "YOUR DATA PATH"
project: "EasySAT"
task: "bump_var_function" # rephase_function, restart_condition ; * to be consistent with your `folder` saving original Solver code..
original: True
original_result: {'time':0,'PAR-2':1} # Your Baseline results (* total sum , not mean)

agent_type: 'advisor-coder-evaluator' # ['advisor-coder-evaluator','coder-evaluator','advisor-coder', 'coder-only']
agent_args_folder: './examples/EasySAT/bump_var_function' # advisor, coder , evaluator PROMPT with their args
temperature: 0.8 # llm-temperature

api_base: 'YOUR　API BASE'
api_key: 'YOUR API KEY'

NeedEval: True
SAT_solver_file_path: './examples/EasySAT/original_EasySAT/EasySAT.cpp'
eval_timeout: 15
results_save_path: './temp/eval_results/'
eval_data_dir: "YOUR TEST DATA PATH"
eval_parallel_size: 36
rand_seed: 42
keep_intermediate_results: False