Instructions on how to load the data?

[blindfish]

Code uses the Cornell movie scripts dataset, but it's unclear how to prepare the .txt files from that dataset to fit the code:

ans = np.load("./datasets/f_idx_a_1.npy")
ques = np.load("./datasets/f_idx_q_1.npy")

with open('./datasets/metadata_1.pkl', 'rb') as f:
        metadata = pickle.load(f)

[purelyvivid]

Hi,
Sorry for late response.
Download the corpus and unzip it.
In data_util/data.py, change the corpus_path to your own corpus path.
Hope that would help. Thanks.