How can I use DPA-2.2.0-v3.0.0b3??

[Hanuklee]

Hello, developers and user

I want to use DPA-2.2.0-v3.0.0b3 pre-trained model. I found "https://www.aissquare.com/models/detail?pageType=models&name=DPA-2.2.0-v3.0.0b3&id=272" this link, and I follow "How to use" section in the link. (Also, I can get the OpenLAM_2.2.0_27heads_beta3.pt)

I get off-line packages of this link (https://github.com/deepmodeling/deepmd-kit/tree/v3.0.0b3), and I successfully installed.
However, I run the follow command,

/home/lhu/src/deepmd-kit/bin/lmp_mpi -i in.lammps

The error lines occur like below.

LAMMPS (2 Aug 2023)
OMP_NUM_THREADS environment is not set. Defaulting to 1 thread. (src/comm.cpp:98)
  using 1 OpenMP thread(s) per MPI task
DeePMD-kit: Successfully load libcudart.so.12
2024-08-29 15:05:17.925720: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2024-08-29 15:05:17.943990: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2024-08-29 15:05:17.944009: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2024-08-29 15:05:17.944500: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
Loaded 1 plugins from /home/lhu/src/deepmd-kit/lib/deepmd_lmp
Reading data file ...
  orthogonal box = (-100.4022 -100 -100.09799) to (100.29419 100.62342 100.52106)
  1 by 1 by 1 MPI processor grid
  reading atoms ...
  55 atoms
  read_data CPU = 0.001 seconds
DeePMD-kit WARNING: Environmental variable DP_INTRA_OP_PARALLELISM_THREADS is not set. Tune DP_INTRA_OP_PARALLELISM_THREADS for the best performance. See https://deepmd.rtfd.io/parallelism/ for more information.
DeePMD-kit WARNING: Environmental variable DP_INTER_OP_PARALLELISM_THREADS is not set. Tune DP_INTER_OP_PARALLELISM_THREADS for the best performance. See https://deepmd.rtfd.io/parallelism/ for more information.
DeePMD-kit WARNING: Environmental variable OMP_NUM_THREADS is not set. Tune OMP_NUM_THREADS for the best performance. See https://deepmd.rtfd.io/parallelism/ for more information.
Summary of lammps deepmd module ...
  >>> Info of deepmd-kit:
  installed to:       /home/lhu/src/deepmd-kit
  source:
  source branch:      HEAD
  source commit:      cbf2de6
  source commit at:   2024-07-27 05:11:58 +0000
  support model ver.: 1.1
  build variant:      cuda
  build with tf inc:  /home/lhu/src/deepmd-kit/lib/python3.11/site-packages/tensorflow/include;/home/lhu/src/deepmd-kit/include
  build with tf lib:  /home/lhu/src/deepmd-kit/lib/python3.11/site-packages/tensorflow/libtensorflow_cc.so.2
  build with pt lib:  torch;torch_library;/home/lhu/src/deepmd-kit/lib/python3.11/site-packages/torch/lib/libc10.so;/home/conda/feedstock_root/build_artifacts/deepmd-kit_1722057349216/_build_env/targets/x86_64-linux/lib/stubs/libcuda.so;/home/lhu/src/deepmd-kit/lib/libnvrtc.so;/home/lhu/src/deepmd-kit/lib/libnvToolsExt.so;/home/lhu/src/deepmd-kit/lib/libcudart.so;/home/lhu/src/deepmd-kit/lib/python3.11/site-packages/torch/lib/libc10_cuda.so
  set tf intra_op_parallelism_threads: 0
  set tf inter_op_parallelism_threads: 0
  >>> Info of lammps module:
  use deepmd-kit at:  /home/lhu/src/deepmd-kitERROR on proc 0: DeePMD-kit C API Error: DeePMD-kit Error: Unsupported model file format (/home/conda/feedstock_root/build_artifacts/deepmd-kit_1722057349216/work/source/lmp/pair_deepmd.cpp:980)
Last command: pair_style deepmd OpenLAM_2.2.0_27heads_beta3.pt
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------

How can I fix it?? Any assistance from someone would be greatly appreciated!!!

[njzjz]

You must use the .pth file obtained with dp --pt freeze.

[caolz]

So using OpenLAM_2.2.0_27heads_beta3.pt directly is not possible right? I need to fine tune this?