add flash_attn_version

Author: yefubao

install/install-requirements.sh

@@ -87,9 +87,10 @@ cu=$(nvcc --version | grep "Cuda compilation tools" | awk '{print $5}' | cut -d
 torch=$(pip show torch | grep Version | awk '{print $2}' | cut -d '+' -f 1 | cut -d '.' -f 1,2)
 cp=$(python3 --version | awk '{print $2}' | awk -F. '{print $1$2}')
 cxx=$(g++ --version | grep 'g++' | awk '{print $3}' | cut -d '.' -f 1)
-wget https://github.com/Dao-AILab/flash-attention/releases/download/v2.8.0.post2/flash_attn-2.8.0.post2+cu${cu}torch${torch}cxx${cxx}abiFALSE-cp${cp}-cp${cp}-linux_x86_64.whl
-pip install --no-cache-dir flash_attn-2.8.0.post2+cu${cu}torch${torch}cxx${cxx}abiFALSE-cp${cp}-cp${cp}-linux_x86_64.whl
-rm flash_attn-2.8.0.post2+cu${cu}torch${torch}cxx${cxx}abiFALSE-cp${cp}-cp${cp}-linux_x86_64.whl
+flash_attn_version="2.8.0.post2"
+wget https://github.com/Dao-AILab/flash-attention/releases/download/v${flash_attn_version}/flash_attn-${flash_attn_version}+cu${cu}torch${torch}cxx${cxx}abiFALSE-cp${cp}-cp${cp}-linux_x86_64.whl
+pip install --no-cache-dir flash_attn-${flash_attn_version}+cu${cu}torch${torch}cxx${cxx}abiFALSE-cp${cp}-cp${cp}-linux_x86_64.whl
+rm flash_attn-${flash_attn_version}+cu${cu}torch${torch}cxx${cxx}abiFALSE-cp${cp}-cp${cp}-linux_x86_64.whl
 
 # From Megatron-LM log
 pip install "git+https://github.com/Dao-AILab/flash-attention.git@v2.7.2#egg=flashattn-hopper&subdirectory=hopper"