fix: clear highres images and reset surya KV cache to prevent memory leak in PdfConverter#1044
Open
aashishkumar-tech wants to merge 1 commit into
Open
Conversation
…fConverter When reusing PdfConverter across multiple PDFs, memory grew unboundedly (~5-6 GB per PDF) due to two issues: 1. highres_image held on every page object after processing completes 2. Surya foundation model KV cache never reset between documents Fixes: datalab-to#1040
Contributor
|
CLA Assistant Lite bot All contributors have signed the CLA ✍️ ✅ |
Author
I have read the CLA document and I hereby sign the CLA |
Author
|
I have read the CLA Document and I hereby sign the CLA |
Author
|
recheck |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
When reusing
PdfConverteracross multiple PDFs in a loop, memory growsunboundedly (~5-6 GB per PDF) and is never reclaimed, as reported in #1040.
Root Cause
Two issues found:
highres_image(192 DPI PIL images) stored on every page object werenever cleared after all processors completed
FoundationPredictor.kv_cachewas never reset between documents,causing torch attention buffers to accumulate
Fix
In
marker/converters/pdf.py, after all processors complete:page.highres_image = Nonefor all pagesmodel.foundation_predictor.kv_cache = Nonefor all surya modelsTesting
Added two regression tests in
tests/converters/test_memory_leak.py:test_highres_images_freed_after_conversion— asserts highres images are None after buildtest_memory_stable_across_multiple_pdfs— asserts <50MB growth across 3 PDFsBoth tests pass ✅
Closes #1040