metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train_100
num_bytes: 76272
num_examples: 100
- name: train_500
num_bytes: 388738
num_examples: 500
- name: train_1k
num_bytes: 782325
num_examples: 1000
- name: train_5k
num_bytes: 3895468
num_examples: 5000
- name: train_full
num_bytes: 8079549
num_examples: 10335
- name: val
num_bytes: 3729778
num_examples: 4576
download_size: 2409420
dataset_size: 16952130
configs:
- config_name: default
data_files:
- split: train_100
path: data/train_100-*
- split: train_500
path: data/train_500-*
- split: train_1k
path: data/train_1k-*
- split: train_5k
path: data/train_5k-*
- split: train_full
path: data/train_full-*
- split: val
path: data/val-*