|
--- |
|
configs: |
|
- config_name: default |
|
data_files: |
|
- split: train |
|
path: data/train-* |
|
dataset_info: |
|
features: |
|
- name: input_ids |
|
sequence: int32 |
|
- name: attention_mask |
|
sequence: int8 |
|
- name: labels |
|
sequence: int64 |
|
splits: |
|
- name: train |
|
num_bytes: 67744337720 |
|
num_examples: 79514 |
|
download_size: 1125510240 |
|
dataset_size: 67744337720 |
|
--- |
|
# Dataset Card for "pg_books-tokenized-bos-eos-chunked-65536" |
|
|
|
The [pg19](https://huggingface.co/datasets/emozilla/pg19) dataset tokenized under LLaMA into 64k chunks, bookended with BOS and EOS |
