When you download a dataset from Hugging Face, the data are stored locally on your computer.
Files from Hugging Face are stored as usual in the huggingface_hub
cache, which is at ~/.cache/huggingface/hub
by default.
See the Hub cache documentation for more details and how to change its location.
The Hub cache allows π€ Datasets to avoid re-downloading dataset files from Hugging Face every time you use them.
π€ Datasets also has its own cache to store datasets converted in Arrow format (the format used by Dataset objects).
This guide focuses on the π€ Datasets cache and will show you how to:
The default π€ Datasets cache directory is ~/.cache/huggingface/datasets
. Change the cache location by setting the shell environment variable, HF_DATASETS_CACHE
to another directory:
$ export HF_DATASETS_CACHE="/path/to/another/directory/datasets"
When you load a dataset, you also have the option to change where the data is cached. Change the cache_dir
parameter to the path you want:
>>> from datasets import load_dataset
>>> dataset = load_dataset('username/dataset', cache_dir="/path/to/another/directory/datasets")
After you download a dataset, control how it is loaded by load_dataset() with the download_mode
parameter. By default, π€ Datasets will reuse a dataset if it exists. But if you need the original dataset without any processing functions applied, re-download the files as shown below:
>>> from datasets import load_dataset
>>> dataset = load_dataset('squad', download_mode='force_redownload')
Refer to DownloadMode for a full list of download modes.
Clean up the Arrow cache files in the directory with Dataset.cleanup_cache_files():
# Returns the number of removed cache files
>>> dataset.cleanup_cache_files()
2
If youβre using a cached file locally, it will automatically reload the dataset with any previous transforms you applied to the dataset. Disable this behavior by setting the argument load_from_cache_file=False
in Dataset.map():
>>> updated_dataset = small_dataset.map(add_prefix, load_from_cache_file=False)
In the example above, π€ Datasets will execute the function add_prefix
over the entire dataset again instead of loading the dataset from its previous state.
Disable caching on a global scale with disable_caching():
>>> from datasets import disable_caching
>>> disable_caching()
When you disable caching, π€ Datasets will no longer reload cached files when applying transforms to datasets. Any transform you apply on your dataset will be need to be reapplied.
If you want to reuse a dataset from scratch, try setting the download_mode
parameter in load_dataset() instead.
Disabling the cache and copying the dataset in-memory will speed up dataset operations. There are two options for copying the dataset in-memory:
Set datasets.config.IN_MEMORY_MAX_SIZE
to a nonzero value (in bytes) that fits in your RAM memory.
Set the environment variable HF_DATASETS_IN_MEMORY_MAX_SIZE
to a nonzero value. Note that the first method takes higher precedence.