python-tensorflowHow can I use Python TensorFlow with Nvidia GPUs?

Using Python TensorFlow with Nvidia GPUs

1. Install the latest version of TensorFlow with GPU support: pip install tensorflow-gpu.
2. Install the latest Nvidia GPU drivers and CUDA Toolkit.
3. Verify your system has a compatible GPU.
4. Configure the GPU environment variables:

export CUDA_VISIBLE_DEVICES=0
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64"

5. Run Python code with TensorFlow-GPU:

import tensorflow as tf
tf.test.is_gpu_available()

# Output
True

6. Train and evaluate models on the GPU.
7. Monitor the GPU usage with the nvidia-smi command line tool.

Helpful links

• TensorFlow GPU Support
• Nvidia Drivers Download
• CUDA Toolkit Download

Edit this code on GitHub