# Install dependencies
git submodule update
git submodule update --init --recursive
CMAKE_ARGS="-DGGML_CUDA=on -DCMAKE_BUILD_PARALLEL_LEVEL=8" uv sync --prerelease=allow  
# might need to specify your CUDA arch:
CMAKE_BUILD_PARALLEL_LEVEL=8 CMAKE_ARGS="-DGGML_CUDA=on -DCMAKE_CUDA_ARCHITECTURES="75"" uv sync --prerelease=allow  
# # if you want to use you metal processor:
CMAKE_ARGS="-DGGML_METAL=on" uv sync --prerelease=allow  
source .venv/bin/activate
# in case you want to edit the llama python bindings
# # if you want to use the GPU:
# CMAKE_BUILD_PARALLEL_LEVEL=8 CMAKE_ARGS="-DGGML_CUDA=on" uv pip install -e llama-cpp-python
# CMAKE_BUILD_PARALLEL_LEVEL=8 CMAKE_ARGS="-DGGML_METAL=on"  uv pip install -e llama-cpp-python
# on a cluster you could start into a interactive environment to compile with nvcc!
srun --gres=gpu -c 12 -u kanben  --time=1:00:00  --pty   bash

# dowload models (make sure to set you HF token!)
tune download NousResearch/Hermes-3-Llama-3.2-3B  --output-dir models/hermes-3-2-3B
huggingface-cli download meta-llama/Llama-3.2-3B original/tokenizer.model --local-dir  models/hermes-3-2-3B
tune download NousResearch/Hermes-3-Llama-3.1-8B  --output-dir models/hermes-3-1-8B
huggingface-cli download meta-llama/Llama-3.1-8B original/tokenizer.model --local-dir  models/hermes-3-1-8B
python manage_models/quantize_all.py