metadata
dataset_info:
- config_name: bpe128000minipile
features:
- name: input_ids
sequence: uint32
- name: uid
dtype: int64
splits:
- name: train
num_bytes: 5641958228
num_examples: 1000000
- name: validation
num_bytes: 2647240
num_examples: 500
- name: test
num_bytes: 55703708
num_examples: 10000
download_size: 2365136824
dataset_size: 5700309176
- config_name: bpe2wp32000minipile
features:
- name: input_ids
sequence: uint16
- name: uid
dtype: int64
splits:
- name: train
num_bytes: 3047712640
num_examples: 1000000
- name: validation
num_bytes: 1419540
num_examples: 500
- name: test
num_bytes: 30008482
num_examples: 10000
download_size: 2938246787
dataset_size: 3079140662
- config_name: bpe32000finewebedu
features:
- name: input_ids
sequence: uint16
- name: uid
dtype: int64
splits:
- name: train
num_bytes: 3615587908
num_examples: 1000000
- name: validation
num_bytes: 1669068
num_examples: 500
- name: test
num_bytes: 35209492
num_examples: 10000
download_size: 3107693507
dataset_size: 3652466468
- config_name: bpe32000minipile
features:
- name: input_ids
sequence: uint16
- name: uid
dtype: int64
splits:
- name: train
num_bytes: 3057036542
num_examples: 1000000
- name: validation
num_bytes: 1423220
num_examples: 500
- name: test
num_bytes: 30093740
num_examples: 10000
download_size: 2948779814
dataset_size: 3088553502
- config_name: bpe8064minipile
features:
- name: input_ids
sequence: uint16
- name: uid
dtype: int64
splits:
- name: train
num_bytes: 3592172606
num_examples: 1000000
- name: validation
num_bytes: 1659704
num_examples: 500
- name: test
num_bytes: 35297914
num_examples: 10000
download_size: 2929205824
dataset_size: 3629130224
- config_name: wordpiece32000minipile
features:
- name: input_ids
sequence: uint16
- name: uid
dtype: int64
splits:
- name: train
num_bytes: 3089305120
num_examples: 1000000
- name: validation
num_bytes: 1435864
num_examples: 500
- name: test
num_bytes: 30398132
num_examples: 10000
download_size: 2976333857
dataset_size: 3121139116
configs:
- config_name: bpe128000minipile
data_files:
- split: train
path: bpe128000minipile/train-*
- split: validation
path: bpe128000minipile/validation-*
- split: test
path: bpe128000minipile/test-*
- config_name: bpe2wp32000minipile
data_files:
- split: train
path: bpe2wp32000minipile/train-*
- split: validation
path: bpe2wp32000minipile/validation-*
- split: test
path: bpe2wp32000minipile/test-*
- config_name: bpe32000finewebedu
data_files:
- split: train
path: bpe32000finewebedu/train-*
- split: validation
path: bpe32000finewebedu/validation-*
- split: test
path: bpe32000finewebedu/test-*
- config_name: bpe32000minipile
data_files:
- split: train
path: bpe32000minipile/train-*
- split: validation
path: bpe32000minipile/validation-*
- split: test
path: bpe32000minipile/test-*
- config_name: bpe8064minipile
data_files:
- split: train
path: bpe8064minipile/train-*
- split: validation
path: bpe8064minipile/validation-*
- split: test
path: bpe8064minipile/test-*
- config_name: wordpiece32000minipile
data_files:
- split: train
path: wordpiece32000minipile/train-*
- split: validation
path: wordpiece32000minipile/validation-*
- split: test
path: wordpiece32000minipile/test-*
Tokenised version of the JeanKaddour/minipile dataset using the tokenisers in repo pietrolesci/tokenisers.