I can't run lammps directly with DPA-1-OC2M model

This is the log file of lammps:
using 1 OpenMP thread(s) per MPI task
plugin load libdeepmd_lmp.so
Loading plugin: deepmd pair style v2.0 by Han Wang
Loading plugin: compute deeptensor/atom v2.0 by Han Wang
Loading plugin: fix dplr v2.0 by Han Wang
units metal
dimension 3
boundary p p p
atom_style atomic
box tilt large
read_data conf.lmp
Reading data file ...
triclinic box = (0.0000000 0.0000000 0.0000000) to (3.2144871 3.2144871 3.2144871) with tilt (0.0000000 0.0000000 0.0000000)
1 by 1 by 1 MPI processor grid
reading atoms ...
2 atoms
read_data CPU = 0.001 seconds
mass 1 55.845
neigh_modify every 1 delay 0 check no
pair_style deepmd OC_10M.pb
Summary of lammps deepmd module ...
>>> Info of deepmd-kit:
installed to: /share/home/02011315/deepmd-kit
source: v2.1.1-1-gdff6aa33
source branch: HEAD
source commit: dff6aa33
source commit at: 2022-05-21 20:41:28 -0400
surpport model ver.:1.1
build float prec: double
build with tf inc: /share/home/02011315/deepmd-kit/include;/share/home/02011315/deepmd-kit/include
build with tf lib: /share/home/02011315/deepmd-kit/lib/libtensorflow_cc.so;/share/home/02011315/deepmd-kit/lib/libtensorflow_framework.so
set tf intra_op_parallelism_threads: 0
set tf inter_op_parallelism_threads: 0
>>> Info of lammps module:
use deepmd-kit at: /share/home/02011315/deepmd-kitERROR: DeePMD-kit Error: TensorFlow Error: NOT_FOUND: Op type not registered 'ProdEnvMatAMix' in binary running on gpu01. Make sure the Op and Kernel are registered in the binary running in this process. Note that if you are loading a saved graph which used ops from tf.contrib, accessing (e.g.) `tf.contrib.resampler` should be done before importing the graph, as contrib ops are lazily registered when the module is first accessed. (/home/conda/feedstock_root/build_artifacts/libdeepmd_1653183660651/work/source/lmp/pair_deepmd.cpp:835)
Last command: pair_style deepmd OC_10M.pb

As I didn't find the version of the DPA-1-OC2M model， I have tried "dp convert-from 0.12" ( and also 1.1, 1.2, 1.3, 2.0) to convert the model to the current 2.1 version, and they all failed. Can you help me to solve this problem?