Datasets:

facebook

/

multilingual_librispeech

like 50

Hello, thanks for the dataset.

I'm trying to get only a part of the dataset but after many tests:

librispeech_test_clean = load_dataset("facebook/multilingual_librispeech", "french", split="unexisting")

This seems to still load a dataset.

I need to load only some part of the datasets like HF dataset documentation show

librispeech_test_clean = load_dataset("facebook/multilingual_librispeech", "french", split="train.1h[:10]")

But it looks like it does not work, still load me around 600 datas.

Do you guys know how can I load only a partial part of the dataset ? Am I misunderstanding HF dataset behaviour ?

Thanks !