[VLM] Create empty videos vector with batch #3129
Conversation
as-suvorov
changed the title
There was a problem hiding this comment.
Pull request overview
Fixes a bug where the generate method created an empty videos vector with a fixed batch size of 1, causing failures when multiple prompts were passed. The fix ensures the empty videos vector is initialized with the same batch size as the number of prompts.
- Updated the
generatemethod to create an empty videos vector sized to match the prompt batch
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
l-bat
left a comment
l-bat
left a comment
There was a problem hiding this comment.
Thanks for fixing this! Now milebench_eval_vlm.py should work with batch > 1 for any VLM model.
There was a problem hiding this comment.
Pull request overview
Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| def idx_to_ov_image_placeholder(idx: int) -> str: | ||
| return f"<ov_genai_image_{idx}>" |
There was a problem hiding this comment.
The function name idx_to_ov_image_placeholder is verbose. Consider renaming it to format_image_placeholder or get_image_placeholder for better readability.
There was a problem hiding this comment.
Pull request overview
Copilot reviewed 2 out of 2 changed files in this pull request and generated no new comments.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
4 participants
Description
Empty videos vector was created with batch 1. It fails if multiple prompts passed. Create empty_videos_vector with prompts batch size
Use ov_genai image placeholders in
milebench_eval_vlm.pysample.CVS-178649
Fixes #(issue)
Checklist: