TensorFlow

TensorFlow (TF) (https://www.tensorflow.org), is the most popular deep learning library. It's developed and maintained by Google. You don't need to explicitly require the use of a GPU; rather TensorFlow will automatically try to use it if you have one. If you have more than one GPU, you must assign operations to each GPU explicitly, or only the first one will be used. To do this, you simply need to type the line that is show in the following code block:

with tensorflow.device("/gpu:1"): 
   # model definition here

Here's an example:

• "/cpu:0": the main CPU of your machine
• "/gpu:0": the first GPU of your machine, if one exists
• "/gpu:1": the second GPU of your machine, if a second exists
• "/gpu:2": the third GPU of your machine, if a third exists, and so on

TensorFlow has a steeper learning curve, compared to the other libraries. You can refer to the TensorFlow documentation to learn how to use it.