ci.yml

prepare:
  steps:
    - name: Clone repository
      command: git clone https://github.com/ggerganov/llama.cpp.git | (cd llama.cpp ;
        git pull)
    - name: Clean
      command: cd llama.cpp && make clean
    - name: Build Llama Cpp
      command: cd llama.cpp && if [ -v NV_LIBCUBLAS_VERSION ]; then echo 'Make for
        gpu' && LLAMA_CUDA=1 make; else echo 'Make for cpu' && make; fi
    - name: Download model
      command: "[ -f
        /home/user/app/llama.cpp/models/Meta-Llama-3-8B-Instruct.Q4_K_M.gguf ]
        || wget -P /home/user/app/llama.cpp/models
        https://huggingface.co/QuantFactory/Meta-Llama-3-8B-Instruct-GGUF/resol\
        ve/main/Meta-Llama-3-8B-Instruct.Q4_K_M.gguf"
test:
  steps: []
run:
  steps:
    - name: Run
      command: cd llama.cpp && if [ -v NV_LIBCUBLAS_VERSION ]; then echo 'Starting gpu
        server' && ./server -m ./models/Meta-Llama-3-8B-Instruct.Q4_K_M.gguf -c
        2048 --port 3000 --host 0.0.0.0 -ngl 35; else echo 'Starting cpu server'
        &&  ./server -m ./models/Meta-Llama-3-8B-Instruct.Q4_K_M.gguf -c 2048
        --port 3000 --host 0.0.0.0; fi