Skip to content

haoran-ni/11667-coding-agent

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

36 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Self-debugging coding agent

A coding agent finetuned from meta-llama/Llama-3.2-1B-Instruct and meta-llama/Llama-3.2-3B-Instruct capable of debugging with self-iterations.

Running

The running tests are done on NERSC. Please make sure cudatoolkit/12.4 is loaded before setting up the conda environment.

conda create --name cmu-llms-final python=3.11
conda activate cmu-llms-final
pip install -r requirements.txt

login wandb
huggingface-cli login

Please note that the environment may be updated frequently, so be sure to check and update your conda environment when there are new commits to the repo.

Fine-tuning

To fine-tune the model, run the following command:

python src/finetuning.py configs/<training_config_file>

Evaluation

To evaluate the model, run the following command:

python src/inference.py configs/<inference_config_file> <name_or_path_to_the_model>

Hugging Face Repository

This project is hosted on Hugging Face under the repository:

Adapters (Better Performance)

Merged Models

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •