01: Test your PyTorch setup¶

Welcome to the laboratory, my eager apprentice; our first incantation is a simple one, to ensure your terminal is ready for the raw power we're about to unleash!

Start with the most basic example of Python.

print("Hello world")

In [1]:

print("Hello world")

print("Hello world")

Hello world

Summoning the Beast!¶

Now, we invoke the great PyTorch itself! Let's check its pulse and see if it has found the precious CUDA cores we so desperately need for our electrifying experiments.

In [2]:

import torch
print(f"PyTorch version: {torch.__version__}")
print(f"CUDA available: {torch.cuda.is_available()}")

import torch print(f"PyTorch version: {torch.__version__}") print(f"CUDA available: {torch.cuda.is_available()}")

PyTorch version: 2.7.0+cu126
CUDA available: True

Behold the Vital Signs!¶

If the stars have aligned and your incantations were correct, you should see a message confirming PyTorch's awakening. But this is merely a surface reading!

PyTorch version: 2.7.0+cu126
CUDA available: True

Now, let us peer deeper into the machine's soul and examine the very essence of its GPU, CuDNN, and other vital components!

In [3]:

#Display a PyTorch version and GPU availability
if torch.cuda.is_available():
    device = torch.device("cuda")          # a CUDA device object
    print(f"Using device: {device}")
    
    # Print CUDA device properties
    print("\nCUDA Device Properties:")
    print(f"Device name: {torch.cuda.get_device_name(0)}")
    print(f"Device properties: {torch.cuda.get_device_properties(0)}")
    print(f"Current device index: {torch.cuda.current_device()}")
    print(f"Device count: {torch.cuda.device_count()}")
    
    # Print CUDA version and capabilities
    print("\nCUDA Information:")
    print(f"CUDA version: {torch.version.cuda}")
    print(f"cuDNN version: {torch.backends.cudnn.version()}")
    print(f"cuDNN enabled: {torch.backends.cudnn.enabled}")
    
    # Print PyTorch memory info
    print("\nPyTorch Memory Information:")
    print(f"Allocated memory: {torch.cuda.memory_allocated(0) / 1024**2:.2f} MB")
    print(f"Cached memory: {torch.cuda.memory_reserved(0) / 1024**2:.2f} MB")
else:
    print('GPU not enabled')
    print("\nPyTorch Information:")
    print(f"PyTorch version: {torch.__version__}")
    print(f"Backend: {torch.get_default_dtype()}")
    

#Display a PyTorch version and GPU availability if torch.cuda.is_available(): device = torch.device("cuda") # a CUDA device object print(f"Using device: {device}") # Print CUDA device properties print("\nCUDA Device Properties:") print(f"Device name: {torch.cuda.get_device_name(0)}") print(f"Device properties: {torch.cuda.get_device_properties(0)}") print(f"Current device index: {torch.cuda.current_device()}") print(f"Device count: {torch.cuda.device_count()}") # Print CUDA version and capabilities print("\nCUDA Information:") print(f"CUDA version: {torch.version.cuda}") print(f"cuDNN version: {torch.backends.cudnn.version()}") print(f"cuDNN enabled: {torch.backends.cudnn.enabled}") # Print PyTorch memory info print("\nPyTorch Memory Information:") print(f"Allocated memory: {torch.cuda.memory_allocated(0) / 1024**2:.2f} MB") print(f"Cached memory: {torch.cuda.memory_reserved(0) / 1024**2:.2f} MB") else: print('GPU not enabled') print("\nPyTorch Information:") print(f"PyTorch version: {torch.__version__}") print(f"Backend: {torch.get_default_dtype()}")

Using device: cuda

CUDA Device Properties:
Device name: NVIDIA GeForce RTX 3080 Laptop GPU
Device properties: _CudaDeviceProperties(name='NVIDIA GeForce RTX 3080 Laptop GPU', major=8, minor=6, total_memory=16383MB, multi_processor_count=48, uuid=4a2f15bc-7268-fcb8-f7b3-9f60002afe35, L2_cache_size=4MB)
Current device index: 0
Device count: 1

CUDA Information:
CUDA version: 12.6
cuDNN version: 90701
cuDNN enabled: True

PyTorch Memory Information:
Allocated memory: 0.00 MB
Cached memory: 0.00 MB

The Apparatus is Ready!¶

Mwahahaha! Excellent! You have successfully interrogated the machine and confirmed that the foundational conduits are in place. The GPU's heart beats strong, and the PyTorch beast is straining at its leash, ready for our command.

With this knowledge, you are one step closer to bending the very fabric of computation to your will! Our instruments are tuned, the lab is humming with potential. Now, the real work begins...