Skip to content

[Test][BugFix] Fix torch.rand usage in triton penalty test #6680

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
zzzzwwjj merged 1 commit into from
Feb 11, 2026
Merged

[Test][BugFix] Fix torch.rand usage in triton penalty test #6680

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 4 additions & 4 deletions tests/e2e/nightly/single_node/ops/singlecard_ops/triton/test_penality.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -127,22 +127,22 @@ def create_test_data(
repetiton_penalty = torch.ones(num_reqs, device=device, dtype=torch.float32)
for i in range(num_reqs):
if torch.rand(1) > 0.3:
repetiton_penalty[i] = torch.rand(1, device).item() * 0.8 + 0.6
repetiton_penalty[i] = torch.rand(1, device=device).item() * 0.8 + 0.6

frequency_penalty = torch.zeros(num_reqs, device=device, dtype=torch.float32)
for i in range(num_reqs):
if torch.rand(1) > 0.5:
frequency_penalty[i] = torch.rand(1, device).item() * 0.2
frequency_penalty[i] = torch.rand(1, device=device).item() * 0.2

presence_penalty = torch.zeros(num_reqs, device=device, dtype=torch.float32)
for i in range(num_reqs):
if torch.rand(1) > 0.5:
presence_penalty[i] = torch.rand(1, device).item() * 0.2
presence_penalty[i] = torch.rand(1, device=device).item() * 0.2

temperature = torch.ones(num_reqs, device=device, dtype=torch.float32)
for i in range(num_reqs):
if torch.rand(1) > 0.2:
presence_penalty[i] = torch.rand(1, device).item() * 1.8 + 0.2
presence_penalty[i] = torch.rand(1, device=device).item() * 1.8 + 0.2
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist Bot Feb 11, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

critical

There appears to be a copy-paste error here. This loop is intended to initialize the temperature tensor, but this line is modifying presence_penalty instead. This will result in incorrect test data, with temperature remaining as a tensor of ones and presence_penalty being potentially overwritten.

Suggested change
presence_penalty[i] = torch.rand(1, device=device).item() * 1.8 + 0.2
temperature[i] = torch.rand(1, device=device).item() * 1.8 + 0.2


idx_mapping = torch.randint(0, num_status, (num_reqs,), device=device, dtype=torch.int32)

Expand Down
Loading