How to use GPU of PyTorch

In this issue, Xiaobian will bring you about how to use PyTorch GPU. The article is rich in content and analyzed and described from a professional perspective. After reading this article, I hope you can gain something.

How GPU is used

Generally we only have one GPU, so training with GPU requires input:

device = torch.device("cuda:0") #Declares that GPU 0 is used

Model = Model(input_size, output_size) # Model is the network name, usually the class name. instantiation Model

model = model.to (device) #Send instantiated model to GPU.

The above step sends the model to the GPU. We also need to send the input data and truth values to the GPU, so we must also have:

inputs, labels = inputs.to(device), labels.to(device)

Note:

In pytorch, inputs.to(device) and inputs.cuda() have the same effect.

In pytorch, b=a.to(device) and a.to(device) are not the same for tensors, the former letting b into the GPU and the latter not letting a into the GPU.

For model, b=a.to(device) and a.to(device) are the same, b and a both go to GPU.

The above is how to use PyTorch GPU shared by Xiaobian for everyone. If you happen to have similar doubts, you may wish to refer to the above analysis for understanding. If you want to know more about it, please pay attention to the industry information channel.