TensorFlow Quick Reference Table – Cheat Sheet.
TensorFlow is a very popular deep-learning library, with its complexity can be overwhelming, especially for new users. Here is a short summary of often-used functions, if you want to download it in pdf it is available here:
TensorFlow CheetSheet – SecretDataScientist.com
If you find it useful please share on social media.
Import TensorFlow: |
|
| import tensorflow as tf | |
Basic math operations: |
|
| tf.add() | sum |
| tf.subtract() | subtraction |
| tf.multiply() | multiplication |
| tf.div() | division |
| tf.mod() | module |
| tf.abs() | absolute value |
| tf.negative() | negative value |
| tf.sign() | return sign |
| tf.reciprocal() | reciprocal |
| tf.square() | square |
| tf.round() | nearest integer |
| tf.sqrt() | square root |
| tf.pow() | power |
| tf.exp() | exponent |
| tf.log() | logarithm |
| tf.maximum() | maximum |
| tf.minimum() | minimum |
| tf.cos() | cosine |
| tf.sin() | sine |
Basic operations on tensors: |
|
| tf.string_to_number() | converts a string to a numeric type |
| tf.cast() | casts to a new type |
| tf.shape() | returns shape of a tensor |
| tf.reshape() | reshapes tensor |
| tf.diag() | creates tensor with given diagonal values |
| tf.zeros() | creates a tensor with all elements set to zero |
| tf.fill() | creates tensor with all elements set given value |
| tf.concat() | concatenates tensors |
| tf.slice() | extracts slice from tensor |
| tf.transpose() | transpose the argument |
| tf.matmul() | matrices multiplication |
| tf.matrix_determinant() | determinant of matrices |
| tf.matrix_inverse() | computes inverse of matrices |
Control Flow: |
|
| tf.while_loop() | repeat body while condition true |
| tf.case() | case operator |
| tf.count_up_to() | incriments ref untill limit |
| tf.tuple() | groups tensors together |
Logical/Comparison Operators: |
|
| tf.equal() | returns truth value element-wise |
| tf.not_equal() | returns truth value of X!=Y |
| tf.less() | returns truth value of X<Y |
| tf.less_equal() | returns truth value of X<=Y |
| tf.greater() | returns truth value of X>Y |
| tf.greater_equal() | returns truth value of X>=Y |
| tf.is_nan() | returns which elements are NaN |
| tf.logical_and() | returns truth value of ‘AND’ for given tensors |
| tf.logical_or() | returns truth value of ‘OR’ for given tensors |
| tf.logical_not() | returns truth value of ‘NOT’ for given tensors |
| tf.logical_xor() | returns truth value of ‘XOR’ for given tensors |
Working with Images: |
|
| tf.image.decode_image() | converts image to tensor type uint8 |
| tf.image.resize_images() | resize images |
| tf.image.resize_image_with_crop_or_pad() | resize image by cropping or padding |
| tf.image.flip_up_down() | flip image horizontally |
| tf.image.rot90() | rotate image 90 degrees counter-clockwise |
| tf.image.rgb_to_grayscale() | converts image from RGB to grayscale |
| tf.image.per_image_standardization() | scales image to zero mean and unit norm |
Neural Networks: |
|
| tf.nn.relu() | rectified linear activation function |
| tf.nn.softmax() | softmax activation function |
| tf.nn.sigmoid() | sigmoid activation function |
| tf.nn.tanh() | hyperbolic tangent activation function |
| tf.nn.dropout | dropout |
| tf.nn.bias_add | adds bias to value |
| tf.nn.all_candidate_sampler() | set of all classes |
| tf.nn.weighted_moments() | returns mean and variance |
| tf.nn.softmax_cross_entropy_with_logits() | softmax cross entropy |
| tf.nn.sigmoid_cross_entropy_with_logits() | sigmoid cross entropy |
| tf.nn.l2_normalize() | normalization using L2 Norm |
| tf.nn.l2_loss() | L2 loss |
| tf.nn.dynamic_rnn() | RNN specified by given cell |
| tf.nn.conv2d() | 2D convolutions given 4D input |
| tf.nn.conv1d() | 1D convolution given 3D input |
| tf.nn.batch_normalization() | batch normalization |
| tf.nn.xw_plus_b() | computes matmul(x,weights)+biases |
High level Machine Learning: |
|
| tf.contrib.keras | Keras API as high level API for TensorFlow |
| tf.contrib.layers.one_hot_column() | one hot encoding |
| tf.contrib.learn.LogisticRegressor() | logistic regression |
| tf.contrib.learn.DNNClassifier() | DNN classifier |
| tf.contrib.learn.DynamicRnnEstimator() | Rnn Estimator |
| tf.contrib.learn.KMeansClustering() | K-Means Clusstering |
| tf.contrib.learn.LinearClassifier() | linear classifier |
| tf.contrib.learn.LinearRegressor() | linear regressor |
| tf.contrib.learn.extract_pandas_data() | extract data from Pandas dataframe |
| tf.contrib.metrics.accuracy() | accuracy |
| tf.contrib.metrics.auc_using_histogram() | AUC |
| tf.contrib.metrics.confusion_matrix() | confusion matrix |
| tf.contrib.metrics.streaming_mean_absolute_error() | mean absolute error |
| tf.contrib.rnn.BasicLSTMCell() | basic lstm cell |
| tf.contrib.rnn.BasicRNNCell() | basic rnn cell |
Placeholders and Variables: |
|
| tf.placeholder() | defines placeholder |
| tf.Variable(tf.random_normal([3, 4], stddev=0.1) | defines variable |
| tf.Variable(tf.zeros([50]), name=’x’) | defines variable |
| tf.global_variables_initializer() | initialize global variables |
| tf.local_variables_initializer() | initialize local variables |
| with tf.device(“/cpu:0”): | pin variable to CPU |
| v = tf.Variable() | |
| with tf.device(“/gpu:0”): | pin variable to GPU |
| v = tf.Variable() | |
| sess = tf.Session() | run session |
| sess.run() | |
| sess.close() | |
| with tf.Session() as session: | run session(2) |
| session.run() | |
| saver=tf.train.Saver() | Saving and restoring variables. |
| saver.save(sess,’file_name’) | |
| saver.restore(sess,’file_name’) | |
Working with Data: |
|
| tf.decode_csv() | converts csv to tensors |
| tf.read_file() | reads file |
| tf.write_file() | writes to file |
| tf.train.batch() | creates batches of tensors |
Was the above useful? Please share with others on social media.
If you want to look for more information, check some free online courses available at coursera.org, edx.org or udemy.com.
Recommended reading list:
  | Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems Through a series of recent breakthroughs, deep learning has boosted the entire field of machine learning. Now, even programmers who know close to nothing about this technology can use simple, efficient tools to implement programs capable of learning from data. This practical book shows you how. By using concrete examples, minimal theory, and two production-ready Python frameworks—scikit-learn and TensorFlow—author Aurélien Géron helps you gain an intuitive understanding of the concepts and tools for building intelligent systems. You’ll learn a range of techniques, starting with simple linear regression and progressing to deep neural networks. With exercises in each chapter to help you apply what you’ve learned, all you need is programming experience to get started. Explore the machine learning landscape, particularly neural nets Use scikit-learn to track an example machine-learning project end-to-end Explore several training models, including support vector machines, decision trees, random forests, and ensemble methods Use the TensorFlow library to build and train neural nets Dive into neural net architectures, including convolutional nets, recurrent nets, and deep reinforcement learning Learn techniques for training and scaling deep neural nets Apply practical code examples without acquiring excessive machine learning theory or algorithm details |