"Invalid argument: Session was not created with a graph before Run()! " when running ~/dp/bin/lmp on HPC cluster

[Daisy315]

Hi,
I am running a case on HPC cluster and it encountered the issue "Invalid argument: Session was not created with a graph before Run()!
/bin/sh: line 1: 3455 Aborted /path/to/my/software/miniconda/envs/deepmd/bin/lmp -i input.lammps -v restart 0"
and this case works successfully on my localhost, so I was wondering how it comes and how to resolve this issue?

And my machine.json is set as below:

"model_devi": [
{
"machine": {
"context_type": "local",
"batch_type": "Slurm",
"local_root":"./",
"remote_root":"/path/to/my/dp-work-dir/"
},
"resources": {
"number_node": 1,
"gpu_per_node": 0,
"cpu_per_node": 4,
"queue_name": "amd_256",
"source_list": [
"~/.bashrc"
],
"module_list": [],
"group_size": 10
},
"command": "/path/to/my/software/miniconda/envs/deepmd/bin/lmp"
}
],