whisper

Whisper

In this directory, you will find examples on how you could apply BigDL-LLM INT4 optimizations on Whisper models on Intel GPUs. For illustration purposes, we utilize the openai/whisper-tiny as a reference Whisper model.

0. Requirements

To run these examples with BigDL-LLM on Intel GPUs, we have some recommended requirements for your machine, please refer to here for more information.

Example: Recognize Tokens using generate() API

In the example recognize.py, we show a basic use case for a Whisper model to conduct transcription using generate() API, with BigDL-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=3.9
conda activate llm
# below command will install intel_extension_for_pytorch==2.1.10+xpu as default
pip install --pre --upgrade bigdl-llm[xpu] -f https://developer.intel.com/ipex-whl-stable-xpu
pip install datasets soundfile librosa # required by audio processing

1.2 Installation on Windows

We suggest using conda to manage environment:

conda create -n llm python=3.9 libuv
conda activate llm
# below command will install intel_extension_for_pytorch==2.1.10+xpu as default
pip install --pre --upgrade bigdl-llm[xpu] -f https://developer.intel.com/ipex-whl-stable-xpu
pip install datasets soundfile librosa # required by audio processing

2. Configures OneAPI environment variables

2.1 Configurations for Linux

source /opt/intel/oneapi/setvars.sh

2.2 Configurations for Windows

call "C:\Program Files (x86)\Intel\oneAPI\setvars.bat"

Note: Please make sure you are using CMD (Anaconda Prompt if using conda) to run the command as PowerShell is not supported.

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

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 ENABLE_SDP_FUSION=1

Note: Please note that libtcmalloc.so can be installed by conda install -c conda-forge -y gperftools=2.10.

3.2 Configurations for Windows

For Intel iGPU

set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1

For Intel Arc™ A300-Series or Pro A60

set SYCL_CACHE_PERSISTENT=1

For other Intel dGPU Series

There is no need to set further environment variables.

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 ./recognize.py --repo-id-or-model-path REPO_ID_OR_MODEL_PATH --repo-id-or-data-path REPO_ID_OR_DATA_PATH --language LANGUAGE

Arguments info:

• --repo-id-or-model-path REPO_ID_OR_MODEL_PATH: argument defining the huggingface repo id for the Whisper model to be downloaded, or the path to the huggingface checkpoint folder. It is default to be 'openai/whisper-tiny'.
• --repo-id-or-data-path REPO_ID_OR_DATA_PATH: argument defining the huggingface repo id for the audio dataset to be downloaded, or the path to the huggingface dataset folder. It is default to be 'hf-internal-testing/librispeech_asr_dummy'.
• --language LANGUAGE: argument defining language to be transcribed. It is default to be english.

Sample Output

openai/whisper-tiny

Inference time: xxxx s
-------------------- Output --------------------
[' Mr. Quilter is the apostle of the middle classes and we are glad to welcome his gospel.']