lammps遇到卡死问题，求助。

懒洋洋喜洋洋

(allegro-pol) root@bohrium-428977-1415901:/personal/b6# ~/lammps/build/lmp -in in.equilibrate
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
Reading data file ...
  orthogonal box = (0 0 0) to (11.898254 11.898254 12.912077)
  1 by 1 by 1 MPI processor grid
  reading atoms ...
  135 atoms
  read_data CPU = 0.003 seconds
Allegro is using input precision f and output precision d
Allegro: Loading model from BaTiO3-E0.nequip.pth
Allegro: Freezing TorchScript model...
Type mapping:
Allegro type | Allegro name | LAMMPS type | LAMMPS name
0 | Ba | 1 | Ba
1 | Ti | 3 | Ti
2 | O | 2 | O
compute allegro will evaluate the quantity polarization of length 3 (src/compute_allegro.cpp:73)
compute allegro will evaluate the quantity polarizability of length 9 (src/compute_allegro.cpp:73)
compute allegro/atom will evaluate the quantity born_charge of length 9 with newton 1 (src/compute_allegro.cpp:63)
Neighbor list info ...
  update: every = 1 steps, delay = 0 steps, check = yes
  max neighbors/atom: 2000, page size: 100000
  master list distance cutoff = 8
  ghost atom cutoff = 8
  binsize = 4, bins = 3 3 4
  1 neighbor lists, perpetual/occasional/extra = 1 0 0
  (1) pair allegro, perpetual
      attributes: full, newton on, ghost
      pair build: full/bin/ghost
      stencil: full/ghost/bin/3d
      bin: standard
Setting up Verlet run ...
  Unit style    : metal
  Current step  : 0
  Time step     : 0.002
Per MPI rank memory allocation (min/avg/max) = 3.929 | 3.929 | 3.929 Mbytes
    PotEng          Fmax          Fnorm          S/CPU         CPULeft   
-100941.72      0.46333129     2.1087595      0              0 我又重新按照之前的文献工作进行，这里使用的pytorch是1.13版本（pip install torch==1.13.0+cu117 torchvision==0.14.0+cu117 torchaudio==0.13.0 --extra-index-url https://download.pytorch.org/whl/cu117）。进行到这一步骤又没有反映了，我查阅AI回答是由于这个问题PyTorch 内部的 autograd 引擎仍然可能启动自己的线程池（例如通过 Intel TBB 或内部的并行后端），这些线程与 LAMMPS 的主线程交互时可能产生死锁。有大佬遇见过类似的问题吗？急急急！！！
复现文献的工作为https://github.com/mir-group/allegro-pol/