Error in compressing a model(v1.2.0)

[Nancy877]

Dear DP developers:
I try to use deepmd-kit v2.0.3 to compress the model trained with v1.2.0 (it has been converted first), command:"dp compress -t compress.json -i frozen.pb -o compress.pb" but the following error is reported:

DEEPMD INFO -----------------------------------------------------------------
DEEPMD INFO training without frame parameter
Traceback (most recent call last):
File "/home/fnancy/soft/deepmd-kit/bin/dp", line 10, in
sys.exit(main())
File "/home/fnancy/soft/deepmd-kit/lib/python3.9/site-packages/deepmd/entrypoints/main.py", line 447, in main
compress(**dict_args)
File "/home/fnancy/soft/deepmd-kit/lib/python3.9/site-packages/deepmd/entrypoints/compress.py", line 119, in compress
train(
File "/home/fnancy/soft/deepmd-kit/lib/python3.9/site-packages/deepmd/entrypoints/train.py", line 103, in train
_do_work(jdata, run_opt, is_compress)
File "/home/fnancy/soft/deepmd-kit/lib/python3.9/site-packages/deepmd/entrypoints/train.py", line 159, in _do_work
model.build(train_data, stop_batch)
File "/home/fnancy/soft/deepmd-kit/lib/python3.9/site-packages/deepmd/train/trainer.py", line 285, in build
self.descrpt.enable_compression(self.model_param['compress']["min_nbor_dist"], self.model_param['compress']['model_file'], self.model_param['compress']['table_config'][0], self.model_param['compress']['table_config'][1], self.model_param['compress']['table_config'][2], self.model_param['compress']['table_config'][3])
File "/home/fnancy/soft/deepmd-kit/lib/python3.9/site-packages/deepmd/descriptor/se_a.py", line 330, in enable_compression
self.table = DPTabulate(
File "/home/fnancy/soft/deepmd-kit/lib/python3.9/site-packages/deepmd/utils/tabulate.py", line 88, in init
self.embedding_net_nodes = get_embedding_net_nodes_from_graph_def(self.graph_def, suffix=self.suffix)
File "/home/fnancy/soft/deepmd-kit/lib/python3.9/site-packages/deepmd/utils/graph.py", line 137, in get_embedding_net_nodes_from_graph_def
assert key.find('bias') > 0 or key.find(
AssertionError: currently, only support weight matrix and bias matrix at the tabulation op!

I wonder if this is because the format or version of the dataset does not match. Could you please give me some advice?

[njzjz]

You need to convert models to v2 version before compressing.
See https://docs.deepmodeling.org/projects/deepmd/en/v2.0.3/freeze/compress.html

[Nancy877]

Thanks for your reply. But I have upgraded the model by "dp convert-from", and the converted model can be run in lammps (v2.0.3), compressing model still has the above problems.

[njzjz]

Could you please provide your model?
cc @denghuilu

[Nancy877]

The model (Sn) is downloaded in DP library "https://dplibrary.deepmd.net//#/project_details?project_id=202011.016", and these are my files
models.zip
:

[njzjz]

The original model enables resnet_dt of the descriptor. Currently, this is not supported by compression.
cc @felix5572: it looks like this model was uploaded by you?