In this directory, you will find examples on how you could use IPEX-LLM optimize_model
API to accelerate Mamba models. For illustration purposes, we utilize the state-spaces/mamba-1.4b and state-spaces/mamba-2.8b as reference Mamba models.
To run these examples with IPEX-LLM on Intel GPUs, we have some recommended requirements for your machine, please refer to here for more information.
In the example generate.py, we show a basic use case for a Mamba model to predict the next N tokens using generate()
API, with IPEX-LLM INT4 optimizations on Intel GPUs.
We suggest using conda to manage environment:
conda create -n llm python=3.11
conda activate llm
# below command will install intel_extension_for_pytorch==2.1.10+xpu as default
pip install --pre --upgrade ipex-llm[xpu] --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/
pip install einops # package required by Mamba
We suggest using conda to manage environment:
conda create -n llm python=3.11 libuv
conda activate llm
# below command will install intel_extension_for_pytorch==2.1.10+xpu as default
pip install --pre --upgrade ipex-llm[xpu] --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/
pip install einops # package required by Mamba
Note
Skip this step if you are running on Windows.
This is a required step on Linux for APT or offline installed oneAPI. Skip this step for PIP-installed oneAPI.
source /opt/intel/oneapi/setvars.sh
For optimal performance, it is recommended to set several environment variables. Please check out the suggestions based on your device.
For Intel Arc™ A-Series Graphics and Intel Data Center GPU Flex Series
export USE_XETLA=OFF
export SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1
export SYCL_CACHE_PERSISTENT=1
For Intel Data Center GPU Max Series
export LD_PRELOAD=${LD_PRELOAD}:${CONDA_PREFIX}/lib/libtcmalloc.so
export SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1
export SYCL_CACHE_PERSISTENT=1
export ENABLE_SDP_FUSION=1
Note: Please note that
libtcmalloc.so
can be installed byconda install -c conda-forge -y gperftools=2.10
.
For Intel iGPU
export SYCL_CACHE_PERSISTENT=1
export BIGDL_LLM_XMX_DISABLED=1
For Intel iGPU
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
For Intel Arc™ A-Series Graphics
set SYCL_CACHE_PERSISTENT=1
Note
For the first time that each model runs on Intel iGPU/Intel Arc™ A300-Series or Pro A60, it may take several minutes to compile.
python ./generate.py
In the example, several arguments can be passed to satisfy your requirements:
--repo-id-or-model-path REPO_ID_OR_MODEL_PATH
: argument defining the huggingface repo id for the Mamba model (e.gstate-spaces/mamba-1.4b
andstate-spaces/mamba-2.8b
) to be downloaded, or the path to the huggingface checkpoint folder. It is default to bestate-spaces/mamba-1.4b
.--tokenizer-repo-id-or-path
: argument defining the huggingface repo id for the tokenizer of Mamba model to be downloaded, or the path to the huggingface checkpoint folder. It is default to beEleutherAI/gpt-neox-20b
.--prompt PROMPT
: argument defining the prompt to be infered (with integrated prompt format for chat). It is default to be'What is AI?'
.--n-predict N_PREDICT
: argument defining the max number of tokens to predict. It is default to be32
.
Inference time: xxxx s
-------------------- Output --------------------
What is AI?
Artificial Intelligence (AI) is a broad term that describes the use of artificial intelligence (AI) to create artificial intelligence (AI). AI is a
Inference time: xxxx s
-------------------- Output --------------------
What is AI?
Artificial Intelligence is a field of study that focuses on creating machines that can perform intelligent functions. These functions are performed by machines that are smarter than humans