Skip to content

[wwb] Update video inputs to avoid big video#3777

Merged
sbalandi merged 2 commits into
openvinotoolkit:masterfrom
sbalandi:fix_video_input
May 13, 2026
Merged

[wwb] Update video inputs to avoid big video#3777
sbalandi merged 2 commits into
openvinotoolkit:masterfrom
sbalandi:fix_video_input

Conversation

@sbalandi
Copy link
Copy Markdown
Contributor

@sbalandi sbalandi commented Apr 28, 2026

Description

The max size of the video in default data for visual-video-text --model-type is 38MB, resolution 2880x1080. Due to big memory loads with qwen2-vl model(CVS-180177) validation is craching on some platforms. We'll limit the video size. Max video size become 6MB. The resolutions of the video in the sample will be: 360x480, 394x480, 640x480 , 640x360 , 480x270 , 480x318 , 1280x720

Set non-default seed to avoid test with small resolution video, task for reason is created - CVS-186452

CVS-185026

Checklist:

  • This PR follows GenAI Contributing guidelines.
  • Tests have been updated or added to cover the new code.
  • This PR fully addresses the ticket.
  • I have made corresponding changes to the documentation.

Copilot AI review requested due to automatic review settings April 28, 2026 15:11
@github-actions github-actions Bot added the category: WWB PR changes WWB label Apr 28, 2026
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

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

Pull request overview

Updates the Who-What Benchmark video default dataset preparation to avoid sampling large videos from the downloaded tar archive, reducing resource usage when running video/VLM evaluations.

Changes:

  • Filters tar archive members by a maximum size threshold before sampling videos.
  • Switches from tar.getnames() to iterating tar.getmembers() to access member sizes.

Comment thread tools/who_what_benchmark/whowhatbench/utils.py Outdated
Comment thread tools/who_what_benchmark/whowhatbench/utils.py
Comment thread tools/who_what_benchmark/whowhatbench/utils.py Outdated
Comment thread tools/who_what_benchmark/whowhatbench/utils.py
@sbalandi sbalandi marked this pull request as draft April 28, 2026 16:55
Copilot AI review requested due to automatic review settings April 28, 2026 16:56
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

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

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 1 comment.

Comment thread tools/who_what_benchmark/whowhatbench/utils.py Outdated
@sbalandi sbalandi marked this pull request as ready for review May 1, 2026 15:02
Copilot AI review requested due to automatic review settings May 1, 2026 15:02
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

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

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 1 comment.

Comment on lines +238 to +239
# max resolution 1280x720, max size 6MB
max_video_size_bytes = 6 * 1024 * 1024
@sbalandi sbalandi requested a review from yatarkan May 1, 2026 15:12
@as-suvorov
Copy link
Copy Markdown
Collaborator

@sbalandi @likholat It seems that tiny-random-llava-next lost it's accuracy on smaller resolutions
https://github.com/openvinotoolkit/openvino.genai/actions/runs/25219360664/job/73949917196?pr=3777#step:6:217
Is it expected?

@as-suvorov as-suvorov self-assigned this May 5, 2026
Copilot AI review requested due to automatic review settings May 8, 2026 18:06
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

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

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.



def run_test(model_id, model_type, optimum_threshold, genai_threshold, tmp_path):
def run_test(model_id, model_type, optimum_threshold, genai_threshold, tmp_path, extra_args=[]):
)
def test_vlm_video(model_id, model_type, tmp_path):
run_test(model_id, model_type, 0.8, 0.8, tmp_path)
# add non default seed to avoid test with small resolution video, reason - CVS-186452
Copilot AI review requested due to automatic review settings May 8, 2026 20:35
@sbalandi sbalandi force-pushed the fix_video_input branch from 1dc921f to 0c2b3dd Compare May 8, 2026 20:35
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

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

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 3 comments.

Comment thread tools/who_what_benchmark/whowhatbench/utils.py
Comment thread tools/who_what_benchmark/whowhatbench/utils.py
Comment thread tools/who_what_benchmark/whowhatbench/utils.py
@sbalandi sbalandi added this pull request to the merge queue May 13, 2026
Merged via the queue into openvinotoolkit:master with commit 496675b May 13, 2026
167 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

category: WWB PR changes WWB

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants