metadata
dataset_info:
features:
- name: prefix
dtype: string
- name: middle
dtype: string
- name: suffix
dtype: string
- name: file_path
dtype: string
- name: repo_name
dtype: string
- name: context
list:
- name: filename
dtype: string
- name: retrieved_chunk
dtype: string
- name: score
dtype: float64
- name: lang
dtype: string
- name: ground_truth
dtype: string
splits:
- name: train
num_bytes: 100624680
num_examples: 9927
download_size: 21775604
dataset_size: 100624680
configs:
- config_name: default
data_files:
- split: train
path: data/train-*