Error with running model CoRGi

[nyongja]

When I run the CoRGi model with the Goodreads dataset, there are several errors.
I made a goodreads dataset with this code, 'research_experiences/GNN/create_goodreads_dataset.py'.

First, in line 325 of the graph_neural_network.py, an assertion occurs because the data format is not GraphData.
So I tried to use the to_graph function in Dataset Class.

And in 842 of the graph_neural_network.py, AttributeError occurs because there is no x attribute in the data.

AttributeError: 'Dataset' object does not have attribute 'x'.

How do I solve this?

[nyongja]

It was solved by changing the dataet_format to gnn_csv! Thank you

[nyongja]

How can I change the sentences.csv to sentences_array.npy in goodreads dataset?

[jyscardioid]

How can I change the sentences.csv to sentences_array.npy in goodreads dataset?

Assuming you already have sentences.csv,
you can run transform_sentence_csv_to_list() in research_experiments/eedi/sentence_to_embedding.py.
It reads the csv file using pd.read_csv and makes it into an array.

[nyongja]

Thank you for your answer 👍