tep_en_fa_para

ds = tfds.load('huggingface:tep_en_fa_para/en-fa')

TEP: Tehran English-Persian parallel corpus. The first free Eng-Per corpus, provided by the Natural Language and Text Processing Laboratory, University of Tehran.

{
    "translation": {
        "languages": [
            "en",
            "fa"
        ],
        "id": null,
        "_type": "Translation"
    }
}