Skip to contents

Create an HFData object from a Hugging Face dataset

Source: R/sugar.R
hfdt.Rd

Creates an HFData object for a dataset hosted on the Hugging Face Hub. The returned object can be converted to an mlr3 task using mlr3::as_task() or accessed through its active bindings.

Usage

hfdt(
  repo_id,
  config = NULL,
  file_name = NULL,
  target = NULL,
  primary_key = NULL,
  split = NULL,
  revision = NULL,
  task_type = c("auto", "classif", "regr"),
  ...
)

Arguments

repo_id

(character(1))
Repository ID of the dataset on the Hugging Face Hub, e.g. "scikit-learn/iris" or "ibm-research/duorc".

config

(character(1)|NULL)
Dataset configuration to load. If NULL, the default configuration is used when available.

file_name

(character(1)|NULL)
Path to a specific dataset file within the repository. This can be used instead of config to load a particular file.

target

(character|NULL)
Name(s) of the target column(s). If NULL, the target must be specified later before converting to an mlr3 task.

primary_key

(character|NULL)
Name(s) of the column(s) that uniquely identify each observation.

split

(character(1)|NULL)
Dataset split to load, such as "train", "test", or "validation". If NULL, all available splits are loaded.

revision

(character(1)|NULL)
Revision of the dataset to load.

task_type

(character(1))
Type of task for the dataset. One of "auto", "classif", or "regr". If "auto", the task type is inferred from the target column(s).

...

Additional arguments passed to HFData.

Value

An HFData object.