This dataset pre-computes all English-centric directions from bible-nlp/biblenlp-corpus, and as a result loading is significantly faster. Loading example:

from datasets import load_dataset dataset = load_dataset("davidstap/biblenlp-corpus-mmteb", "eng-arb", trust_remote_code=True) dataset DatasetDict({ train: Dataset({ features: ['eng', 'arb'], num_rows: 28723 }) validation: Dataset({ features: ['eng', 'arb'], num_rows: 1578 })… See the full description on the dataset page: https://huggingface.co/datasets/davidstap/biblenlp-corpus-mmteb.