Unrecognized pair style 'deepmd'

[EricGaoOm]

I was running md in DPA2, an warning occured like this
I was curious about such warning caused by incompatible version of lammps and deepmd-kit or other reasons

[EricGaoOm]

ERROR: Unrecognized pair style 'deepmd' (src/force.cpp:275)
Last command: pair_style deepmd dpa2_model.pth

[njzjz]

Could you list how you install the program, how you run the program, and the entire output of the program?

[96XinHe]

I'm experiencing the same problem. I downloaded the offline packages and installed them via the sh command (both vesion 3.0.0b4 and 3.0.0rc0). Subsequent simulations using lammps gave the same error, whether using the se_a2 I had previously trained with TF in the old deepmd-kit or the newly trained DPA-2 model:

LAMMPS (29 Aug 2024)
OMP_NUM_THREADS environment is not set. Defaulting to 1 thread. (src/comm.cpp:98)
  using 1 OpenMP thread(s) per MPI task
Open of file /data/run01/scw6dyq/software/DP_series/deepmd-kit_3.0.0rc0/lib/deepmd_lmp/dpplugin.so failed: /lib64/libc.so.6: version `GLIBC_2.25' not found (required by /data/run01/scw6dyq/software/DP_series/deepmd-kit_3.0.0rc0/bin/../lib/./python3.10/site-packages/tensorflow/../../.././libcares.so.2)
Loaded 0 plugins from /data/run01/scw6dyq/software/DP_series/deepmd-kit_3.0.0rc0/lib/deepmd_lmp
Reading data file ...
  triclinic box = (0 0 0) to (59.0364 51.127022 55.6378) with tilt (-29.5182 0 0)
WARNING: Triclinic box skew is large. LAMMPS will run inefficiently. (src/domain.cpp:221)
  1 by 1 by 1 MPI processor grid
  reading atom labelmap ...
  reading atoms ...
  15876 atoms
  read_data CPU = 0.049 seconds
ERROR: Unrecognized pair style 'deepmd' (src/force.cpp:275)
Last command: pair_style      deepmd freeze_model.pth

If I use an older version of deepmd-kit 2.2.11 that I installed using the same method, I can read the .pb model and run the simulation just fine.

I notice the erro information:
version 'GLIBC_2.25' not found (required by /data/run01/scw6dyq/software/DP_series/deepmd-kit_3.0.0rc0/bin/../lib/./python3.10/site-packages/tensorflow/../../.././libcares.so.2)
and i checked my version by
ldd --version

ldd (GNU libc) 2.17
Copyright (C) 2012 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Written by Roland McGrath and Ulrich Drepper.

It looks like the gnu version is too low on the platform I'm using right? Is it possible to include the gnu version in the next update package?

[njzjz]

Related to conda-forge/c-ares-feedstock#38 (comment)

[njzjz]

You can downgrade c-ares by yourself.

The next version of the offline packages will ship an old version of c-ares.

[96XinHe]

Thanks, I’ve successfully downgraded c-ares from version 1.34.3 to 1.32.3 using the following commands:

conda config --add channels conda-forge
conda config --set channel_priority strict
conda install c-ares=1.32.3

Now, the lmp command no longer reports an error.

[EricGaoOm]

Thanks, I will try this solution!