ayo/ipex-llm
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

Edit gpu doc (#9583)

Browse source 
* harness: run llama2-7b

* harness: run llama2-7b

* harness: run llama2-7b

* harness: run llama2-7b

* edit-gpu-doc

* fix some format problem

* fix spelling problems

* fix evaluation yml

* delete redundant space

* fix some problems

* address comments

* change link
This commit is contained in:
Yining Wang 2023-12-11 14:59:07 +08:00 committed by GitHub
parent 45721f3473
commit a04a027b4c
1 changed files with 38 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

39
docs/readthedocs/source/doc/LLM/Overview/install_gpu.md
View file

@ -5,7 +5,7 @@
Install BigDL-LLM for GPU supports using pip through:
```bash
pip install --pre --upgrade bigdl-llm[xpu] -f https://developer.intel.com/ipex-whl-stable-xpu # install bigdl-llm for GPU
pip install --pre --upgrade bigdl-llm[xpu] -f https://developer.intel.com/ipex-whl-stable-xpu # install bigdl-llm for Intel GPU
```
Please refer to [Environment Setup](#environment-setup) for more information.
@ -15,6 +15,11 @@ Please refer to [Environment Setup](#environment-setup) for more information.
The above command will install ``intel_extension_for_pytorch==2.0.110+xpu`` as default. You can install specific ``ipex``/``torch`` version for your need.
.. important::
Currently ``https://developer.intel.com/ipex-whl-stable-xpu`` is the only achievable source for ``-f`` option since IPEX 2.0.110+xpu and corresponding torch versions are not released on pypi.
.. important::
``bigdl-llm`` is tested with Python 3.9, which is recommended for best practices.
@ -68,6 +73,8 @@ conda activate llm
pip install --pre --upgrade bigdl-llm[xpu] -f https://developer.intel.com/ipex-whl-stable-xpu # install bigdl-llm for GPU
```
If you encounter network issues when installing ipex, you can refer to [Installing bigdl-llm[xpu] dependencies from Wheels](#installing-bigdl-llm[xpu]-dependencies-from-wheels) as an alternative method.
Then for running a LLM model with BigDL-LLM optimizations, several environment variables are recommended:
```bash
@ -77,3 +84,33 @@ source /opt/intel/oneapi/setvars.sh
export USE_XETLA=OFF
export SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1
```
## Installing bigdl-llm[xpu] dependencies from Wheels
You can also install BigDL-LLM dependencies for Intel XPU from source achieves. First you need to install the target torch/torchvision/ipex versions from downloaded whls [here](http://ec2-52-27-27-201.us-west-2.compute.amazonaws.com/ipex-release.php?device=xpu&repo=us&release=stable) before installing bigdl-llm.
```bash
# get the wheels on Linux system
wget https://intel-extension-for-pytorch.s3.amazonaws.com/ipex_stable/xpu/torch-2.0.1a0%2Bcxx11.abi-cp39-cp39-linux_x86_64.whl
wget https://intel-extension-for-pytorch.s3.amazonaws.com/ipex_stable/xpu/torchvision-0.15.2a0%2Bcxx11.abi-cp39-cp39-linux_x86_64.whl
wget https://intel-extension-for-pytorch.s3.amazonaws.com/ipex_stable/xpu/intel_extension_for_pytorch-2.0.110%2Bxpu-cp39-cp39-linux_x86_64.whl
```
```eval_rst
.. note::
Compatible versions of ``torch`` and ``torchvision`` are prerequisites for installing ``intel_extension_for_pytorch``.
```
Then you may install directly from the wheel archives using following commands:
```bash
# install the packages from the wheels
python -m pip install torch-2.0.1a0+cxx11.abi-cp39-cp39-linux_x86_64.whl
python -m pip install torchvision-0.15.2a0+cxx11.abi-cp39-cp39-linux_x86_64.whl
python -m pip install intel_extension_for_pytorch-2.0.110+xpu-cp39-cp39-linux_x86_64.whl
# install bigdl-llm for Intel GPU
pip install --pre --upgrade bigdl-llm[xpu]
```