cd Quark/examples/torch/language_modeling/llm_ptq/
export exclude_layers="lm_head *self_attn* *mlp.gate"
python3 quantize_quark.py --model_dir $MODEL_DIR \
                          --quant_scheme mxfp4 \
                          --num_calib_data 128 \
                          --exclude_layers $exclude_layers \
                          --skip_evaluation \
                          --multi_gpu  \
                          --trust_remote_code \
                          --model_export hf_format \
                          --output_dir $output_dir

# Reinstall vLLM
pip uninstall vllm -y
git clone https://github.com/vllm-project/vllm.git
cd vllm
git checkout v0.13.0
pip install -r requirements/rocm.txt
python setup.py develop
cd ..

VLLM_ROCM_USE_AITER=1 \
VLLM_DISABLE_COMPILE_CACHE=1 \
vllm serve "$MODEL" \
    --tensor-parallel-size 4 \
    --trust-remote-code \
    --max-model-len 32768 \
    --port 8899