Efficient data loader for text dataset using torch.utils.data.Dataset, collate_fn and torch.utils.data.DataLoader.
$ git clone https://github.com/yunjey/seq2seq-dataloader.git
$ cd seq2seq-dataloader
$ pip install nltk
$ python
$ import nltk
$ nltk.download('punkt')
$ python build_vocab.py
For usage, please see example.ipynb.