Run finish_chat() before start_chat() in VLM pipeline. #2470
Conversation
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
github-actions
bot
added
category: visual language
Wovchena
approved these changes
Jul 22, 2025
There was a problem hiding this comment.
Pull Request Overview
This PR ensures that finish_chat() is called before start_chat() in both VLM and Continuous Batching pipelines to prevent accumulation of history images when multiple start_chat() calls are made. This addresses ticket CVS-170106.
- Adds
finish_chat()calls beforestart_chat()in both pipeline implementations - Removes direct state reset code from VLM pipeline implementation, delegating cleanup to
finish_chat() - Removes
m_image_idreset from continuous batching base class
Reviewed Changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
| src/cpp/src/visual_language/pipeline.cpp | Adds finish_chat() call before start_chat() and removes inline state reset logic |
| src/cpp/src/continuous_batching/pipeline_base.cpp | Removes m_image_id reset from base implementation |
| src/cpp/src/continuous_batching/pipeline.cpp | Adds finish_chat() call before start_chat() |
| if (!m_inputs_embedder->get_kv_cache_state().get_state().empty()) { | ||
| m_history.clear(); | ||
| } | ||
| m_inputs_embedder->start_chat(system_message); |
There was a problem hiding this comment.
With the removal of state reset logic, ensure that finish_chat() properly handles all the cleanup that was previously done here, including m_image_id reset, state reset, attention_mask reshaping, and history clearing.
Copilot uses AI. Check for mistakes.
rkazants
reviewed
Jul 23, 2025
rkazants
approved these changes
Jul 23, 2025
Merged
via the queue into
openvinotoolkit:master
with commit Jul 25, 2025
c7d56b4
98 of 101 checks passed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Run finish_chat() before start_chat() in VLM pipeline and CB pipeline to prevent accumulating of history images when multiple start_chat() are used.
Ticket: CVS-170106