Warn at startup when the video content safety classifier is disabled - #171
Merged
Conversation
Implements option 1 from NVIDIA#170. create_video_guardrail_runner builds a runner whose safety_models list is empty (VideoContentSafetyFilter is commented out pending false-positive fixes), so enabling guardrails gives text checks and face blur with nothing in the logs distinguishing "checked and safe" from "not checked". This adds a single warning at runner construction naming the effective coverage. The warning is guarded on the list being empty, so restoring VideoContentSafetyFilter silences it automatically. Signed-off-by: Stew Tong <stewarttong@gmail.com>
lfengad
approved these changes
Aug 10, 2026
lfengad
enabled auto-merge (squash)
August 12, 2026 06:06
Dinghow
approved these changes
Aug 12, 2026
foreverlms
approved these changes
Aug 12, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Implements option 1 from #170, as invited there.
create_video_guardrail_runnercurrently builds a runner whosesafety_modelslist is empty (theVideoContentSafetyFilterentry is commented out pending false-positive fixes), so a user who enables guardrails gets text checks and face blur but no video content screening, with nothing in the logs distinguishing "checked and safe" from "not checked" until the generic per-request line, which reads as a safe verdict.This adds a single startup warning at runner construction when the video safety-model list is empty:
The warning is guarded on the list being empty, so restoring
VideoContentSafetyFiltersilences it automatically with no further change. No behavior changes otherwise; the per-request logic incore.pyis untouched.