Installation and GPU setup in TensorFlow - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is the first step to install TensorFlow with GPU support?

The first step is to install the correct NVIDIA GPU drivers for your graphics card to enable GPU acceleration.

beginner

Why do you need CUDA and cuDNN for TensorFlow GPU setup?

CUDA is a platform that allows TensorFlow to use NVIDIA GPUs for computation. cuDNN is a library that speeds up deep learning operations on GPUs.

beginner

How can you check if TensorFlow is using the GPU after installation?

You can run `tf.config.list_physical_devices('GPU')` in Python. If it returns a list with GPU devices, TensorFlow is using the GPU.

beginner

What command installs TensorFlow with GPU support using pip?

Use `pip install tensorflow` as the main package now includes GPU support by default in recent versions.

intermediate

What is a common issue if TensorFlow does not detect your GPU after setup?

Common issues include incompatible CUDA/cuDNN versions, missing drivers, or environment variables not set correctly.

Which driver must be installed first to enable TensorFlow GPU support?

A NVIDIA GPU driver

B Intel CPU driver

C AMD GPU driver

D USB driver

TensorFlow GPU support requires NVIDIA GPU drivers because it uses CUDA technology.

What is the role of cuDNN in TensorFlow GPU setup?

A It updates the operating system

B It installs Python packages

C It provides deep learning acceleration libraries

D It manages CPU cores

cuDNN is a GPU-accelerated library for deep neural networks that speeds up TensorFlow computations.

How do you verify if TensorFlow detects your GPU?

A Run pip list

B Run tf.config.list_physical_devices('GPU') in Python

C Check the CPU usage in Task Manager

D Restart the computer

This TensorFlow command lists all detected GPU devices.

Which package do you install to get TensorFlow with GPU support in recent versions?

A tensorflow

B tensorflow-gpu

C tensorflow-cpu

D tensorflow-lite

The main tensorflow package now includes GPU support by default.

What might cause TensorFlow not to detect your GPU after installation?

A Too much RAM

B No internet connection

C Wrong Python version

D Incompatible CUDA or cuDNN versions

Using wrong versions of CUDA or cuDNN is a common cause of GPU detection failure.

Explain the steps to set up TensorFlow to use your GPU for faster training.

Think about what software and libraries are needed for GPU acceleration.

Describe how you can check if TensorFlow is actually using your GPU after installation.

Focus on TensorFlow commands that list hardware devices.