metadata
language:
- kk
dataset_info:
features:
- name: query
dtype: string
- name: label
dtype: int64
splits:
- name: train
num_bytes: 48132276.72226
num_examples: 299986
- name: validation
num_bytes: 3971985.30456
num_examples: 24997
- name: test
num_bytes: 4117527.57144
num_examples: 24998
download_size: 29750787
dataset_size: 56221789.59826
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
The dataset contains a few None values for the query column. Please clean the dataset after the load_dataset util:
dataset = load_dataset("nlp-team-issai/rag-decision-dataset-kk")
def remove_na_rows(example):
return example["query"] is not None and example["label"] is not None
cleaned_dataset = DatasetDict({
split: dataset[split].filter(remove_na_rows)
for split in dataset
})