metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
dataset_info:
features:
- name: label
dtype: float64
- name: dataset_name
dtype: string
- name: input_ids
sequence: int32
- name: token_type_ids
sequence: int8
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 232895922
num_examples: 949728
- name: validation
num_bytes: 17255970
num_examples: 69711
- name: test
num_bytes: 96102951
num_examples: 425205
download_size: 123150665
dataset_size: 346254843
Original Dataset from: https://huggingface.co/datasets/glue
This dataset is adapted from https://huggingface.co/datasets/gmongaras/BERT_Base_Cased_512_GLUE
Every split besides the ax split is in this dataset.
Lines above 512 tokens from the BERT-cased (bert-base-cased) tokenizer are removed in the original dataset
If in any case the sentences are longer than 512 tokens, they are subsetted.
Original labels and dataset categories are retained.