Inoob/SentenceFragmentsDataset

Sentence Fragment Dataset

Labeled dataset for sentence fragment text classification derived from aclImdb_v1 with both positive and negative comments.

This is in encoded .h5 file.

Data Structure

The structure is as below:

{
  "label": [...],
  "input_ids": [...],
  "attention_mask": [...]
}

label - int (0: sentence fragment, 1: whole sentence)

input_ids - torch.tensor (tokenized version of each sentence fragment in pytorch tensors)

attention_mask - torch.tensor (attention mask return value of each sentence fragment in pytorch tensors)

NOTE: TO SEE THE TEXT, GO TO MY OTHER DATASET. HOWEVER, THIS DATASET WILL NOT LOAD THE INPUT ID AND ATTENTION MASK CORRECTLY INTO PYTORCH TENSORS.

Loading the Daaset

git clone https://huggingface.co/datasets/Inoob/SentenceFragmentsDataset

If that failed, try

git-lfs clone https://huggingface.co/datasets/Inoob/SentenceFragmentsDataset

Then, load it using h5py:

train split:

import h5py
import torch

with h5py.File('./SentenceFragmentsDataset/train.h5', 'r') as hf:
    label = list(hf['label'])
    
    input_ids = [torch.tensor(data) for data in hf['input_ids']]
    
    attention_mask = [torch.tensor(data) for data in hf['attention_mask']]

data_train={"label":label, "input_ids": input_ids, "attention_mask":attention_mask}

test split:

import h5py
import torch

with h5py.File('./SentenceFragmentsDataset/test.h5', 'r') as hf:
    label = list(hf['label'])
    
    input_ids = [torch.tensor(data) for data in hf['input_ids']]
    
    attention_mask = [torch.tensor(data) for data in hf['attention_mask']]

data_test={"label":label, "input_ids": input_ids, "attention_mask":attention_mask}