Loading a single split with `streaming=False` triggers download of all files

#8
by marcospiau - opened

Hi,

I am trying to download a single config name and split (pt-validation) and the amount of data is small enough to be downloaded and loaded in memory. When I use ds = datasets.load_dataset('mc4', 'pt', split='validation', streaming=True), the only files downloaded are the ones I am interested, but when I use stream=False, a progress bar indicates that all the 1024 training files are being downloaded and prepared.

Below is the code excerpt responsible for this:

    def _split_generators(self, dl_manager):
        data_urls = {}
        for split in ["train", "validation"]:
            data_urls[split] = [
                _DATA_URL.format(
                    language=lang,
                    split_suffix="-validation" if split == "validation" else "",
                    index=index,
                    n_shards=_N_SHARDS_PER_SPLIT[lang][split],
                )
                for lang in self.config.languages
                for index in range(_N_SHARDS_PER_SPLIT[lang][split])
            ]
        train_downloaded_files = dl_manager.download(data_urls["train"])
        validation_downloaded_files = dl_manager.download(data_urls["validation"])
        return [
            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepaths": train_downloaded_files}),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION, gen_kwargs={"filepaths": validation_downloaded_files}
            ),
        ]

Is there a way to change this behavior to download only the data I am interested instead of all files?

Thanks,
Marcos

Sign up or log in to comment