In this directory, you will find examples on how you could apply IPEX-LLM INT4 optimizations on Qwen models on Intel GPUs. For illustration purposes, we utilize the Qwen-7B-Chat as a reference Qwen model.
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 Qwen 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 "transformers<4.37.0"
pip install tiktoken einops transformers_stream_generator # additional package required for Qwen-7B-Chat to conduct generation
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 "transformers<4.37.0"
pip install tiktoken einops transformers_stream_generator # additional package required for Qwen-7B-Chat to conduct generation
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
For Intel iGPU and 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 --repo-id-or-model-path REPO_ID_OR_MODEL_PATH --prompt PROMPT --n-predict N_PREDICT
Arguments info:
--repo-id-or-model-path REPO_ID_OR_MODEL_PATH
: argument defining the huggingface repo id for the Qwen model (e.gQwen/Qwen-7B-Chat
) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be'Qwen/Qwen-7B-Chat'
.--prompt PROMPT
: argument defining the prompt to be infered (with integrated prompt format for chat). It is default to be'AI是什么?'
.--n-predict N_PREDICT
: argument defining the max number of tokens to predict. It is default to be32
.
Inference time: xxxx s
-------------------- Prompt --------------------
<human>AI是什么? <bot>
-------------------- Output --------------------
<human>AI是什么? <bot>AI,即人工智能,是指计算机科学的一个分支,它企图创造能够完成任务的智能机器,这些任务通常需要人类智能才能完成。
Inference time: xxxx s
-------------------- Prompt --------------------
<human>What is AI? <bot>
-------------------- Output --------------------
<human>What is AI? <bot>AI, or artificial intelligence, refers to the ability of a machine or computer program to perform tasks that typically require human intelligence, such as visual perception, speech recognition