mirrors/docling
1
0
Fork 0
mirror of https://github.com/DS4SD/docling.git synced 2025-12-08 12:48:28 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix page_range stopping at page 32 by using dynamic batch_size

Browse Source 
Co-authored-by: cau-git <60343111+cau-git@users.noreply.github.com>
This commit is contained in:
copilot-swe-agent[bot]
2025-11-19 13:25:00 +00:00
parent 18f705b235
commit 58fc6ccf86
2 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
docling/experimental/pipeline/threaded_layout_vlm_pipeline.py
View File

@@ -252,7 +252,7 @@ class ThreadedLayoutVlmPipeline(BasePipeline):
proc = ProcessingResult(total_expected=total_pages) proc = ProcessingResult(total_expected=total_pages)
fed_idx = 0 fed_idx = 0
batch_size = 32 batch_size = total_pages
try: try:
while proc.success_count + proc.failure_count < total_pages: while proc.success_count + proc.failure_count < total_pages:

2
docling/pipeline/standard_pdf_pipeline.py
View File

@@ -545,7 +545,7 @@ class StandardPdfPipeline(ConvertPipeline):
proc = ProcessingResult(total_expected=total_pages) proc = ProcessingResult(total_expected=total_pages)
fed_idx: int = 0 # number of pages successfully queued fed_idx: int = 0 # number of pages successfully queued
batch_size: int = 32 # drain chunk batch_size: int = total_pages # drain chunk
try: try:
while proc.success_count + proc.failure_count < total_pages: while proc.success_count + proc.failure_count < total_pages:
# 1) feed - try to enqueue until the first queue is full # 1) feed - try to enqueue until the first queue is full