metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: code
dtype: string
- name: repo_path
dtype: string
splits:
- name: train
num_bytes: 88302798272
num_examples: 13368477
download_size: 1680002223
dataset_size: 88302798272
license: bigcode-openrail-m
Dataset Card for "starcoder_filtered"
A version of the starcoder dataset filtered based on data quality. Data was labeled with a rater model, and low-ranking rows were removed.