dp_frz

[k50112113]

I encountered this error when I tried to freeze my trained model:

from ._conv import register_converters as _register_converters
Traceback (most recent call last):
File "/home/scl6/project-beckman/deepmd-kit_r0.11/build/bin/dp_frz", line 81, in
freeze_graph(args.folder, args.output, args.nodes)
File "/home/scl6/project-beckman/deepmd-kit_r0.11/build/bin/dp_frz", line 30, in freeze_graph
input_checkpoint = checkpoint.model_checkpoint_path
AttributeError: 'NoneType' object has no attribute 'model_checkpoint_path'

I didn't face any error when I was training my model, so my question is:

1. It seems like a tensorflow problem, does this mean that I didn't install my tensorflow correctly?
2. Does dp_frz depends on tensorflow of python interface only?

[njzjz]

It seems that tensorflow cannot find your checkpoint. For the second question, yes.

[kaijiz]

could anyone tell me how to fix this problem? what is the meaning of checkpoint?

[njzjz]

A file named checkpoint. If it doesn't exist, something went wrong with dp train.