imdb

Basic Support for Large Movie Review Dataset

http://ai.stanford.edu/~amaas/data/sentiment/

Requirements

• Python 3.6 >=
• Numpy

data_loader

import data_loader or from . import data_loader Note: import statement depends on your source root

data_loader.load_train_data(train_dir_path, valid_rate=0.1)

Input

• train_dir_path: training directory path e.g. './aclImdb/train/'
• valid_rate: validation data rate (0 - 1), 0.1 by default

Output

• train_texts: training text list
• train_labels: training label array (numpy.ndarray, 0: negative, 1:positive)
• valid_texts: validation text list
• valid_labels: validation label array (numpy.ndarray, 0: negative, 1:positive)

---

data_loader.load_test_data(test_dir_path)

Input

• test_dir_path: test directory path e.g. './aclImdb/test/'

Output

• test_texts: test text list
• test_labels: test label array (numpy.ndarray, 0: negative, 1:positive)

---

data_loader.load_vocab_dict(vocab_file_path)

Input

• vocab_file_path: vocabulary file path e.g. './aclImdb/imdb.vocab'

Output

• vocab_dict: vocabulary dictionary (key: str, value: int)