ARQ

Setup

To set up the environment for this project, follow these steps:

1. Install Conda: Ensure you have Conda installed on your system. Refer to the following link for an installation guide. https://conda.io/projects/conda/en/latest/user-guide/install/index.html.

2. Create Environment and Install Dependencies:

   • Create a new Conda environment:

     conda create --name your_env_name python=3.8
     conda activate your_env_name

   • Install PyTorch and related packages with CUDA 11.3 support:

     pip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu113

   • Install the remaining dependencies listed in requirements.txt:

     pip install -r requirements.txt

Training the Initial Model

To train the initial model, use the following command as an example:

bash run/pretrain_resnet20.sh

Running Policy Search

To perform the policy search, use the following command as an example:

bash run/search_resnet20.sh

Evaluation

To evaluate the quantization policy, follow these steps:

1. Replace the policy in rl_quantize.py with the policy obtained from the search phase.

2. Run the evaluation script like:

   bash run/eval_resnet20.sh