 myhloli
|
d7d85a2881
feat(ocr): implement language-specific OCR processing
|
7 月之前 |
|
myhloli
|
cf4ea78dac
refactor: remove torchtext deprecation warning handling
|
8 月之前 |
|
myhloli
|
af27c0cc81
refactor(magic_pdf): support mps device and optimize image processing
|
8 月之前 |
|
myhloli
|
1b34f7e4ff
refactor(magic_pdf): replace PIL with NumPy for image processing
|
8 月之前 |
|
myhloli
|
f5112e2157
fix(model): move environment variable settings to global scope
|
9 月之前 |
|
myhloli
|
79c8a5c8cb
feat(table): upgrade RapidTable to1.0.3 and add sub-model support
|
10 月之前 |
|
myhloli
|
f350222614
feat(model): improve batch analysis logic and support npu
|
10 月之前 |
|
myhloli
|
c20e9a1e84
feat(layout): improve title block handling and layout detection
|
10 月之前 |
|
myhloli
|
203b8f9004
fix(device): enable MPS support and fix related issues
|
10 月之前 |
|
myhloli
|
50f4841716
refactor(device): optimize memory cleaning and device selection
|
10 月之前 |
|
myhloli
|
7990e7dfbb
feat(model): add npu support and optimize table model
|
10 月之前 |
|
myhloli
|
9e4ebea939
refactor(magic_pdf): remove YOLO_VERBOSE setting and update YOLOv8 prediction verbosity
|
11 月之前 |
|
myhloli
|
6a75d7dce5
perf(layout): optimize layout detection for PDF extraction
|
11 月之前 |
|
myhloli
|
f5d812b313
feat(layout): improve layout detection for DocLayout_YOLO model
|
11 月之前 |
|
myhloli
|
012a46e07d
refactor(magic-pdf): optimize model initialization and concurrency control
|
11 月之前 |
|
myhloli
|
47a83d28f5
refactor(ocr): replace AtomModelSingleton with ocr_model_init for OCR model instantiation
|
11 月之前 |
|
myhloli
|
f2a92d5782
refactor(model): implement thread-safe OCR model initialization
|
11 月之前 |
|
myhloli
|
a1744b770f
refactor(magic_pdf): remove unused threading lock and model initialization code
|
11 月之前 |
|
myhloli
|
488660dddd
refactor(model): replace AtomModelSingleton with ocr_model_init for OCR model initialization
|
11 月之前 |
|
myhloli
|
6f636b6e7e
refactor(model): replace ModelSingleton with direct model initialization and improve threading
|
11 月之前 |
|
赵小蒙
|
878f3de004
refactor(magic_pdf): optimize model initialization and threading
|
11 月之前 |
|
myhloli
|
f1e2f08467
refactor(model): move page total time logging to custom model analysis
|
1 年之前 |
|
myhloli
|
e52bd023f7
Merge remote-tracking branch 'origin/dev' into dev
|
1 年之前 |
|
myhloli
|
b2e37a2d1b
feat(ocr): improve text detection and OCR accuracy
|
1 年之前 |
 icecraft
|
b492c19c4c
refactor: move some constants or enums defs to config folder
|
1 年之前 |
|
myhloli
|
08f46125a0
refactor(model): rename and restructure model modules
|
1 年之前 |
|
myhloli
|
fe2c2c0d8e
feat(table): add RapidOCR support for RapidTable model
|
1 年之前 |
|
myhloli
|
240fe99e3c
feat(table): integrate RapidTable model for table recognition
|
1 年之前 |
|
myhloli
|
4b0f11769d
refactor(model): remove unused code and simplify OCR model initialization
|
1 年之前 |
|
myhloli
|
11f23843b1
feat(table): upgrade StructEqTable model and integrate into PDF Extract Kit
|
1 年之前 |
