TensorFlow Python Cheatsheet

TensorFlow is an open-source machine learning framework developed by the Google Brain team. It is widely used for building and training machine learning models, especially deep learning models. To help you navigate the powerful features of TensorFlow more efficiently, we’ve put together a cheatsheet that serves as a quick reference guide for both beginners and experienced practitioners.


Before diving into TensorFlow, make sure you have it installed. You can install it using pip:

pip install tensorflow

For GPU support, you can install the GPU version:

pip install tensorflow-gpu

Importing TensorFlow:

import tensorflow as tf


TensorFlow is built around the concept of tensors, which are n-dimensional arrays. Understanding tensors is fundamental to working with TensorFlow.

Creating Tensors:

# Scalar (0-dimensional tensor)
scalar_tensor = tf.constant(5)

# Vector (1-dimensional tensor)
vector_tensor = tf.constant([1, 2, 3])

# Matrix (2-dimensional tensor)
matrix_tensor = tf.constant([[1, 2, 3], [4, 5, 6]])

# Tensor (n-dimensional tensor)
tensor = tf.constant([[[1, 2, 3], [4, 5, 6]], [[7, 8, 9], [10, 11, 12]]])


Variables are mutable tensors. They are used to store and update model parameters during training.

# Creating a variable
variable = tf.Variable(initial_value=[1, 2, 3])

# Updating a variable
variable.assign([4, 5, 6])


TensorFlow allows you to perform various mathematical operations on tensors.

# Element-wise addition
result = tf.add(tensor1, tensor2)

# Matrix multiplication
result = tf.matmul(matrix1, matrix2)

# Element-wise multiplication
result = tf.multiply(tensor1, tensor2)

Neural Networks:

TensorFlow excels in building and training neural networks. Here’s a simple example:

# Define a simple neural network
model = tf.keras.Sequential([
    tf.keras.layers.Dense(units=128, activation='relu', input_shape=(input_size,)),
    tf.keras.layers.Dense(units=10, activation='softmax')

# Compile the model
model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])

# Train the model
model.fit(x_train, y_train, epochs=10, batch_size=32)

Saving and Loading Models:

Saving and loading models is crucial for reusing trained models.

# Save model

# Load model
loaded_model = tf.keras.models.load_model('my_model.h5')

TensorFlow Lite:

TensorFlow Lite is a lightweight version of TensorFlow for mobile and edge devices.

# Convert model to TensorFlow Lite format
converter = tf.lite.TFLiteConverter.from_keras_model(model)
tflite_model = converter.convert()

# Save TensorFlow Lite model
with open('model.tflite', 'wb') as f:

GPU Acceleration:

Take advantage of GPU acceleration for faster training.

# Check for GPU availability
print("Num GPUs Available: ", len(tf.config.experimental.list_physical_devices('GPU')))

This cheatsheet provides a glimpse into the essential aspects of TensorFlow, but there’s much more to explore. TensorFlow’s versatility makes it suitable for a wide range of machine learning tasks, from image classification to natural language processing. As you delve deeper into the world of TensorFlow, refer to this cheatsheet to streamline your workflow and make the most of this powerful framework.


1. What is TensorFlow, and how is it different from other machine learning frameworks?

TensorFlow is an open-source machine learning framework developed by Google. It is designed to facilitate the development and training of machine learning models, particularly deep neural networks. TensorFlow provides a comprehensive ecosystem of tools, libraries, and community support. Its computational graph paradigm allows for efficient execution on CPUs and GPUs. TensorFlow distinguishes itself with its flexibility, scalability, and wide adoption in both research and industry.

2. How can I check if my TensorFlow installation is utilizing GPU acceleration?

To check if your TensorFlow installation is utilizing GPU acceleration, you can use the following code snippet:
import tensorflow as tf # Check for GPU availability print("Num GPUs Available: ", len(tf.config.experimental.list_physical_devices('GPU')))
This will print the number of available GPUs. If it’s greater than zero, TensorFlow is configured to use GPU acceleration.

3. What is the difference between TensorFlow 1.x and TensorFlow 2.x?

TensorFlow 2.x introduced several improvements and changes compared to TensorFlow 1.x. One significant difference is the eager execution mode in TensorFlow 2.x, which allows for immediate evaluation of operations. This makes the framework more intuitive and similar to Python’s NumPy library. Additionally, TensorFlow 2.x integrates the Keras high-level API, making it the official high-level API for model development. The transition to TensorFlow 2.x is encouraged due to its improved usability and enhanced features.

4. How do I handle model saving and loading in TensorFlow?

Saving and loading models in TensorFlow is straightforward. To save a model:
To load the saved model:
loaded_model = tf.keras.models.load_model('my_model.h5')
This allows for easy reuse of trained models for inference or further training.

5. What is TensorFlow Lite, and when should I use it?

TensorFlow Lite is a lightweight version of TensorFlow designed for mobile and edge devices. It optimizes models for deployment on resource-constrained platforms. You should use TensorFlow Lite when you need to deploy machine learning models on mobile devices or embedded systems where computational resources are limited. The framework allows for the conversion of TensorFlow models to a format suitable for deployment on these devices, balancing performance and resource efficiency.