update norm computation for upcoming PyTorch update

[namgyu-youn]

Replaces torch.norm() with torch.linalg.vector_norm() for upcoming PyTorch update.

Based on PyTorch - docs, torch.norm() will no longer be used

Following is the test code:

import torch

# Create test tensors
torch.manual_seed(42)
X = torch.randn(10, 10)
X_hat, A_reg = torch.randn(10, 10), torch.randn(5, 5)

print("X shape:", X.shape)
print("A_reg shape:", A_reg.shape)
print()

# Old Version (torch.norm)
print("=== torch.norm ===")
relative_error = torch.dist(X, X_hat, p=torch.inf) / torch.norm(X, p=torch.inf)
norm_result = torch.norm(A_reg, p=torch.inf)

print("Relative error:", relative_error.item())
print("A_reg infinity norm:", norm_result.item())
print()

# New Version (torch.linalg.vector_norm)
print("=== (torch.linalg.vector_norm ===")
relative_error = torch.dist(X, X_hat, p=torch.inf) / torch.linalg.vector_norm(X, ord=torch.inf)
norm_result = torch.linalg.vector_norm(A_reg, ord=torch.inf)

print("Relative error:", relative_error.item())
print("A_reg infinity norm:", norm_result.item())

And the result is:

X shape: torch.Size([10, 10])
A_reg shape: torch.Size([5, 5])

=== torch.norm ===
Relative error: 1.512144684791565
A_reg infinity norm: 1.9775909185409546

=== torch.linalg.vector_norm ===
Relative error: 1.512144684791565
A_reg infinity norm: 1.9775909185409546

[tsunghsienlee]

Great! Really appreciated that you look-ahead for this!

[facebook-github-bot]

@tsunghsienlee has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@tsunghsienlee merged this pull request in abb683d.