mistral

Mistral

In this directory, you will find examples on how you could apply IPEX-LLM INT4 optimizations on Mistral models on Intel GPUs. For illustration purposes, we utilize the mistralai/Mistral-7B-Instruct-v0.1 and mistralai/Mistral-7B-v0.1 as reference Mistral models.

Requirements

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.

Example: Predict Tokens using generate() API

In the example generate.py, we show a basic use case for a Mistral model to predict the next N tokens using generate() API, with IPEX-LLM INT4 optimizations on Intel GPUs.

1. Install

1.1 Installation on Linux

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/

1.2 Installation on Windows

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/

2. Configures OneAPI environment variables for Linux

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

3. Runtime Configurations

For optimal performance, it is recommended to set several environment variables. Please check out the suggestions based on your device.

3.1 Configurations for Linux

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 by conda install -c conda-forge -y gperftools=2.10.

For Intel iGPU

export SYCL_CACHE_PERSISTENT=1
export BIGDL_LLM_XMX_DISABLED=1

3.2 Configurations for Windows

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.

4. Running examples

python ./generate.py --prompt 'What is AI?'

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 Mistral model (e.g. mistralai/Mistral-7B-Instruct-v0.1 and mistralai/Mistral-7B-v0.1) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be 'mistralai/Mistral-7B-Instruct-v0.1'.
• --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 be 32.

2.3 Sample Output

mistralai/Mistral-7B-Instruct-v0.1

Inference time: xxxx s
-------------------- Output --------------------
[INST] What is AI? [/INST] AI stands for Artificial Intelligence. It is a branch of computer science that focuses on the development of intelligent machines that work, react, and even think like humans

mistralai/Mistral-7B-v0.1

Inference time: xxxx s
-------------------- Output --------------------
[INST] What is AI? [/INST]

[INST] Artificial Intelligence (AI) is a branch of computer science that deals with the simulation of intelligent behavior in computers. It is a broad

2.4 Deepspeed Tensor Parallel Inference

For platforms without sufficient GPU memory to serve Mistral on one signle GPU card, Deepspeed Tensor Parallel can help to scale the inference to multiple cards, which enables large model and increases resource utilization. Please to refer to here.