The cpu or gpu version is installed using the conda :
conda install deepmd-kit=*=cpu lammps-dp==cpu -c deepmodeling or
conda install deepmd-kit==gpu lammps-dp==*gpu -c deepmodeling,
then I installed the lammp-dp python interface by using : conda install -c deepmodeling pylammps-dp;
then installed the phonolammps using :pip install phonolammps
However,the calculated phonon spectrums are very different!It seems the CPU Version give the correct result,while the GPU version is not.
attached is files used for calculations.
I wondered whether i did mistakes in my calculations? or it's a bug?
DP.zip