Skip to content
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.

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

Fix divide by zero and unit conversion for throughput display #500

Merged
merged 2 commits into from
Mar 8, 2024

Conversation

nv-hwoo
Copy link
Contributor

@nv-hwoo nv-hwoo commented Mar 8, 2024

  • When calculating ITL, set number of token to default to 1 when the response is empty (for now)
  • Convert latency units from nanosecond to second in order to avoid near-zero throughput metrics.

@nv-hwoo nv-hwoo requested review from tgerdesnv and dyastremsky March 8, 2024 00:00
@tgerdesnv tgerdesnv merged commit 58b2a74 into feature-genai-pa Mar 8, 2024
3 checks passed
@tgerdesnv tgerdesnv deleted the hwoo-fix-llm-metrics branch March 8, 2024 12:25
debermudez pushed a commit that referenced this pull request Mar 12, 2024
* Fix divide by zero

* Convert ns to secs for throughput metrics
debermudez pushed a commit that referenced this pull request Mar 13, 2024
* Fix divide by zero

* Convert ns to secs for throughput metrics
mc-nv pushed a commit that referenced this pull request Mar 13, 2024
* Fix divide by zero

* Convert ns to secs for throughput metrics
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Development

Successfully merging this pull request may close these issues.

3 participants