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

fix deepspeed dependencies to install (#9400)

Browse source 
* remove reductant parameter from deepspeed install

* Update install.sh

* Update install.sh
This commit is contained in:
Heyang Sun 2023-11-13 16:42:50 +08:00 committed by GitHub
parent 8c603014ac
commit da6bbc8c11
1 changed files with 17 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

22
python/llm/example/CPU/Deepspeed-AutoTP/install.sh
View file

@ -1,9 +1,21 @@
#!/bin/bash
# install torch
# 1. install oneccl for intel mpi
# can skip this step if oneccl/oneapi is already installed on your machine
# report to https://github.com/oneapi-src/oneCCL if any issue
git clone https://github.com/oneapi-src/oneCCL.git
cd oneCCL
mkdir build
cd build
cmake ..
make -j install
mkdir -p /opt/intel/oneccl
mv ./build/_install/env /opt/intel/oneccl
# 2. install torch and ipex
pip install torch==2.1.0
# install torchccl
pip install intel_extension_for_pytorch==2.1.0 -f https://developer.intel.com/ipex-whl-stable-cpu
# install torchccl (oneccl binding for pytorch)
pip install https://intel-extension-for-pytorch.s3.amazonaws.com/torch_ccl/cpu/oneccl_bind_pt-2.1.0%2Bcpu-cp39-cp39-linux_x86_64.whl
# install deepspeed
# 3. install deepspeed
pip install deepspeed==0.11.1
# exclude intel deepspeed extension, which is only for XPU
pip uninstall intel-extension-for-deepspeed --ignore-missing
# 4. exclude intel deepspeed extension, which is only for XPU
pip uninstall intel-extension-for-deepspeed