experiments.py: error: argument --distmult_state_dict_path: expected one argument

[swxhha]

When I train an RL model (policy gradient + reward shaping) on the NELL-995 dataset, I execute the following commands:
./experiment-rs.sh configs/nell-995-rs.sh --train 0 --test
Then I ran into the following problem:
experiments.py: error: argument --distmult_state_dict_path: expected one argument
What is the role of distmult_state_dict_path ? What input does it need?

[kingsaint]

@swxhha Sorry for the delayed response.
You first need to train a ConvE model. As I've described in the README note To train the RL models using reward shaping, make sure 1) you have pre-trained the embedding-based ConvE model and 2) set the file path pointers conve_state_dict_path to the pre-trained embedding-based models correctly in the configs/<dataset>-rs.sh or configs/<dataset>.sh files.