 myhloli
|
dda878377b
|
feat: add character bounding box metrics and script role classification for improved text rendering
|
2026-06-04 02:05:03 +08:00 |
|
|
myhloli
|
2cf7a42453
|
refactor: unify OCR inference functions by consolidating detection and recognition calls
|
2026-06-04 01:39:22 +08:00 |
|
|
myhloli
|
f79b4356c3
|
feat: add functionality to skip broken PDF pages during rewrite process
|
2026-06-02 19:34:32 +08:00 |
|
|
myhloli
|
33ebaa3868
|
refactor: simplify inference locking mechanism and update docstrings
|
2026-06-02 17:51:32 +08:00 |
|
|
myhloli
|
86b84b9f27
|
fix: ensure proper closure of PDFium child objects to prevent resource leaks
|
2026-06-02 17:30:01 +08:00 |
|
|
myhloli
|
949b1b9223
|
fix: ensure proper closure of PDFium child objects to prevent resource leaks
|
2026-06-02 17:01:49 +08:00 |
|
|
myhloli
|
7e0877b71b
|
Merge remote-tracking branch 'origin/dev' into dev
|
2026-06-02 15:52:04 +08:00 |
|
|
myhloli
|
95a7c5c290
|
fix: enhance PDF processing by ensuring proper resource management and error handling
|
2026-06-02 15:51:45 +08:00 |
|
|
myhloli
|
88d881d299
|
fix: increase maximum PDF render processes to enhance rendering performance
|
2026-06-02 15:33:53 +08:00 |
|
|
myhloli
|
2d7485547a
|
fix: replace global inference lock with specific locks for Layout, MFR, and OCR processing
|
2026-06-02 03:26:17 +08:00 |
|
|
myhloli
|
b169958beb
|
fix: reduce maximum PDF render processes to improve resource management
|
2026-06-02 03:24:43 +08:00 |
|
|
myhloli
|
728e0794e2
|
fix: refactor memory cleaning functions to simplify implementation and avoid circular dependencies
|
2026-06-02 03:18:18 +08:00 |
|
|
myhloli
|
b068c34dc0
|
fix: implement global inference lock for serializing Layout/MFR/OCR processing
|
2026-06-02 02:45:19 +08:00 |
|
|
myhloli
|
aa0bda48be
|
fix: remove memory cleaning condition for post-processing in middle JSON conversion
|
2026-06-02 02:28:09 +08:00 |
|
|
myhloli
|
fca67ac4a9
|
fix: remove memory cleaning condition for post-processing in middle JSON conversion
|
2026-06-02 02:27:38 +08:00 |
|
|
myhloli
|
c65e95a443
|
fix: implement inference admission lock and quiet zone for improved concurrency management
|
2026-06-02 02:22:55 +08:00 |
|
|
myhloli
|
8b4c1d3683
|
fix: refactor inference calls to use dedicated run functions for layout, MFR, and OCR detection/recognition
|
2026-06-02 00:03:57 +08:00 |
|
|
myhloli
|
7a385e7e94
|
fix: enhance CID font usage detection and normalize font name handling in PDF processing
|
2026-06-01 20:10:23 +08:00 |
|
|
myhloli
|
02d757fffa
|
fix: switch PDF rendering to use 'spawn' method for multiprocessing and improve task submission handling
|
2026-06-01 17:30:36 +08:00 |
|
|
myhloli
|
bd60d158dc
|
fix: update external link to the latest model version
|
2026-06-01 10:55:38 +08:00 |
|
|
myhloli
|
11e845b3d0
|
fix: improve error handling during PDF render executor shutdown
|
2026-06-01 10:52:56 +08:00 |
|
|
myhloli
|
05b2414959
|
chore: log title leveling execution duration
|
2026-05-29 19:08:58 +08:00 |
|
|
myhloli
|
2d3fb59c67
|
fix: keep model output for client-side output generation
|
2026-05-29 17:22:41 +08:00 |
|
|
myhloli
|
c9f223d7fa
|
doc(cli): support boolean client-side output generation option
|
2026-05-29 00:25:39 +08:00 |
|
|
Xiaomeng Zhao
|
88a8c317c1
|
Merge pull request #5042 from opendatalab/master
master->dev
|
2026-05-28 23:57:59 +08:00 |
|
|
myhloli
|
06b5ff1522
|
Update version.py with new version
|
2026-05-28 15:56:13 +00:00 |
|
|
Xiaomeng Zhao
|
9697823ad0
|
Merge pull request #5040 from opendatalab/dev
3.2.1
mineru-3.2.1-released
|
2026-05-28 19:37:08 +08:00 |
|
|
Xiaomeng Zhao
|
8ad4b713d7
|
Merge pull request #5039 from myhloli/dev
Dev
|
2026-05-28 19:34:39 +08:00 |
|
|
myhloli
|
583a9ee6ac
|
chore: update vLLM Docker base image to 0.21.0
|
2026-05-28 18:12:30 +08:00 |
|
|
myhloli
|
5fe1fd28c7
|
refactor: remove scipy dependency from unet table line ordering and add unit tests
|
2026-05-28 15:46:55 +08:00 |
|
|
myhloli
|
4bae02c295
|
feat: refactor finalize_middle_json to support server-side post-processing and enhance title leveling
|
2026-05-27 23:19:24 +08:00 |
|
|
myhloli
|
0acd7903ee
|
feat: refactor finalize_middle_json to support server-side post-processing and enhance title leveling
|
2026-05-27 22:26:15 +08:00 |
|
|
myhloli
|
cdca3b5709
|
feat: refactor finalize_middle_json to support server-side post-processing and enhance title leveling
|
2026-05-27 22:22:24 +08:00 |
|
|
myhloli
|
ac1795a60c
|
feat: refactor finalize_middle_json to support server-side post-processing and enhance title leveling
|
2026-05-27 22:22:13 +08:00 |
|
|
myhloli
|
6ec3af8157
|
feat: refactor finalize_middle_json to support server-side post-processing and enhance title leveling
|
2026-05-27 22:18:16 +08:00 |
|
|
myhloli
|
381e1612c8
|
feat: refactor finalize_middle_json to support server-side post-processing and enhance title leveling
|
2026-05-27 22:14:26 +08:00 |
|
|
myhloli
|
2bdf60409e
|
feat: refactor finalize_middle_json to support server-side post-processing and enhance title leveling
|
2026-05-27 22:13:00 +08:00 |
|
|
myhloli
|
bd8d33af82
|
feat: refactor finalize_middle_json to support server-side post-processing and enhance title leveling
|
2026-05-27 22:11:55 +08:00 |
|
|
myhloli
|
2add3d60af
|
feat: refactor finalize_middle_json to support server-side post-processing and enhance title leveling
|
2026-05-27 21:33:02 +08:00 |
|
|
myhloli
|
1c2ff68660
|
feat: refactor finalize_middle_json to support server-side post-processing and enhance title leveling
|
2026-05-27 20:48:56 +08:00 |
|
|
myhloli
|
baa7274a0f
|
feat: add client-side output generation option for middle JSON finalization
|
2026-05-27 20:47:04 +08:00 |
|
|
myhloli
|
f51547c9d2
|
feat: streamline middle JSON validation for client-side output generation
|
2026-05-27 20:45:29 +08:00 |
|
|
myhloli
|
6823cca9b7
|
feat: improve title line average height calculation by validating bbox data
|
2026-05-27 20:25:56 +08:00 |
|
|
myhloli
|
5e03d57908
|
feat: add client-side output generation option to defer markdown/content-list generation
|
2026-05-27 19:59:41 +08:00 |
|
|
myhloli
|
e1d6401cff
|
feat: enhance title line average height resolution and ensure completeness for LLM processing
|
2026-05-27 18:47:57 +08:00 |
|
|
myhloli
|
4808896aea
|
feat: update title average line height calculation for consistency in finalized middle JSON
|
2026-05-27 18:36:35 +08:00 |
|
|
myhloli
|
800c61b073
|
feat: refactor title line height calculation and simplify logic for average height resolution
|
2026-05-27 18:14:46 +08:00 |
|
|
myhloli
|
311e3279ca
|
feat: rename client-side title leveling option to output generation and implement regeneration of client-side outputs
|
2026-05-27 17:07:27 +08:00 |
|
|
myhloli
|
33af492456
|
feat: add client-side title leveling option and integrate regeneration of title leveled outputs
|
2026-05-27 16:42:15 +08:00 |
|
|
myhloli
|
4ce0ef703a
|
feat: implement title leveling normalization and enhance role restoration in PDF processing
|
2026-05-27 16:11:59 +08:00 |
|