Weekly Progress of DeePMD-kit 20220912

[HuangJiameng]

Important:

Recently, some problems with model inference of Python interface in DeePMD-kit 2.1.0-2.1.3 CPU version are found. If the model is trained with the 2.1.0-2.1.3 CPU version, due to the model's flaws, the MD process may generate bad data, such as data with abnormally large forces. This issue was fixed in version 2.1.4. Please note that 2.1.0-2.1.3 GPU version is not affected. Besides, LAMMPS in 2.1.0-2.1.3 CPU/GPU version is not affected, which means if one uses LAMMPS in DeePMD-kit 2.1.0 CPU version to run MD process with a normal model, then everything is all right.

What's new

• fix DeprecationWarning of imp #1896
• remove get_platform in setup.py #1897
• bump to C++ 17 for TF 2.10 #1898
• compare converted lr with almost equal #1901
• Fix bugs when init_frz_model using tebd. #1891
• add tutorial and publication links to docs #1904

Issue

• [BUG] LAMMPS installed with in DeePMD-kit 2.1.4 offline package cannot recognized pair style 'deepmd' #1903
• ERROR: Could not build wheels for deepmd-kit, which is required to install pyproject.toml-based projects #1905
• Get USER-DEEPMD without building DeepMD-kit source code #1906

Discussion

• deepmd-kit package is inconsitent with TensorFlow Runtime #1869
  You can try to use the following command CONDA_OVERRIDE_CUDA=11.6 pwd/miniconda3/bin/conda create -n deepmd deepmd-kit=*=*gpu -c https://conda.deepmodeling.org/ -y if you encounter RuntimeError: This deepmd-kit package is inconsistent with TensorFlow Runtime...
• rcut while running the molecular dynamics simulations using trained potential #1902
  The potential is designed to be continuous at the rcut point.
• Generating Input Data #1877
• Trying parallel training for cpu version of deepmd-kit #1900
  One needs to install horovod to make mpirun work.
• Can the potential function of carbide high-entropy ceramics be trained? #1908
• Is it possible to run single precision MD with easy-installed deepmd and LAMMPS? #1907