# Getting PyTorch to Run on the GPU

PyTorch supports GPU acceleration, which can significantly speed up computations. Here’s how to use the GPU in PyTorch:

Before using the GPU, ensure that it’s available:

import torch

# Check if GPU is available
if torch.cuda.is_available():
    print("GPU is available.")
else:
    print("GPU is not available.")

You can move tensors to the GPU using the .to() method or .cuda() method. Here’s an example:

import torch

# Create a tensor and move it to the GPU
tensor_cpu = torch.tensor([1.0, 2.0, 3.0])
tensor_gpu = tensor_cpu.to('cuda')  # Move tensor to GPU

print("Tensor on GPU:", tensor_gpu)

Explanation:

• .to('cuda'): Moves the tensor to the GPU.
• .cuda(): Another way to move a tensor to the GPU.

After computations are done on the GPU, you might want to move tensors back to the CPU:

import torch

# Create a tensor on the GPU
tensor_gpu = torch.tensor([1.0, 2.0, 3.0]).to('cuda')

# Move the tensor back to the CPU
tensor_cpu = tensor_gpu.to('cpu')
print("Tensor on CPU:", tensor_cpu)

Explanation:

• .to('cpu'): Moves the tensor back to the CPU.

Let’s review the key points

1. Check GPU Availability: Use torch.cuda.is_available() to see if a GPU is available.
2. Move Tensors to GPU: Use .to('cuda') to move tensors or models to the GPU.
3. Move Tensors Back to CPU: Use .to('cpu') to move tensors back to the CPU.