Wiki_dpr.py error (ID mismatch between lines {id} and vector {vec_id}

🤗Datasets
1

I’m trying to run a rag example, and the dataset is wiki_dpr.
wiki_dpr download and extracting have been completed successfully.
However, at the generating train split stage, an error from wiki_dpr.py keeps popping up.

Especially in “_generate_examples” :

  1. The following error occurs in the line id, text, title = line.strip().split(“\t”)
    ValueError: not enough values ​​to unpack (expected 3, got 2)
    → This part handles exceptions so that even if an error occurs, it passes.
  2. ID mismatch between lines {id} and vector {vec_id}
    This error seems to occur at the line " assert int(id) == int(vec_id),".
    After I handled the exception in the split error, generating train split progressed to 80%, but an id mismatch error occurred at about the 16200000th vector id.

Debugging is even more difficult because it takes a long time to download and split wiki_dpr. I need help. thank you in advance!!
version: python 3.8, and others are referenced from requirements.txt.