# CoACD Parameter Optimization with RL

This repository uses reinforcement learning to find optimal parameters for CoACD convex decomposition on a given mesh.

## Quick Start

1. **Clone the repository**

   ```bash
   git clone https://github.com/LeonLiu4/coacd.git --recursive
   cd coacd

2. Create and activate the environment

   conda env create -f environment.yml
   conda activate coacd_env

3. Install PyTorch3D

   • Linux

     pip install "git+https://github.com/facebookresearch/pytorch3d.git@stable"

   • macOS

     export KMP_DUPLICATE_LIB_OK=TRUE
     pip install --no-build-isolation --no-deps \
       git+https://github.com/facebookresearch/pytorch3d.git@stable
     unset KMP_DUPLICATE_LIB_OK

4. Run training

   # CPU only
   CUDA_VISIBLE_DEVICES= python -m coacd_ppo_train --device cpu
   
   # GPU
   python -m coacd_ppo_train