metadata
dataset_info:
features:
- name: tokenizer_id
dtype: int64
- name: hash
dtype: string
- name: vocab_hash
dtype: string
- name: vocab_size
dtype: int64
- name: model_type
dtype: string
- name: num_merges
dtype: int64
- name: has_normalizer
dtype: bool
- name: has_pre_tokenizer
dtype: bool
- name: has_post_processor
dtype: bool
- name: has_decoder
dtype: bool
- name: num_added_tokens
dtype: int64
- name: normalizer_type
dtype: string
- name: pre_tokenizer_type
dtype: string
- name: decoder_type
dtype: string
- name: normalizer_types
list: string
- name: pre_tokenizer_types
list: string
- name: decoder_types
list: string
- name: version
dtype: string
- name: truncation
dtype: string
- name: padding
dtype: string
- name: added_tokens
dtype: string
- name: normalizer
dtype: string
- name: pre_tokenizer
dtype: string
- name: post_processor
dtype: string
- name: decoder
dtype: string
- name: model
dtype: string
splits:
- name: train
num_bytes: 75252466551
num_examples: 24798
download_size: 40124756481
dataset_size: 75252466551
configs:
- config_name: default
data_files:
- split: train
path: data/train-*