Tensorflow-quickstart-offline

Base on Tensorflow official quickstart, but to load train data locally(offline) instead of downloading online, avoid blocking by GFW in China or others.

For whom

Blocked by GFW(like China) can't download tensorflow.keras.datasets.* by calling load_data() methods.

Method

keras.datasets	local_datasets
imdb.get_word_index	get_word_index
imdb.load_data	load_data_imdb
mnist.load_data	load_data_mnist
fashion_mnist.load_data	load_data_fashion_mnist
boston_housing.load_data	load_data_boston_housing

Usage

For quick start

Open notebook in official_quickstart.

For your examples project

copy utils directory to any project.

python

from utils.local_datasets import load_data_fashion_mnist

(train_images, train_labels), (test_images, test_labels) = load_data_fashion_mnist()

jupyter notebook

import os
import sys
# add utils path to sys.path in order to import
# PS: be aware of the absolute path of utils
module_path = os.path.abspath(os.path.join('utils'))
if module_path not in sys.path:
    sys.path.append(module_path)
    
from local_datasets import load_data_fashion_mnist

(train_images, train_labels), (test_images, test_labels) = load_data_fashion_mnist()