Yesterday
-
Model card: narrower model table, figures open full size on click
-
Model card: full documentation with real-document examples, architecture diagrams and corrected scores
3 days ago
-
Add recogniser v3 (crnn_m_v3_owner): v2.3 fine-tuned on real Khmer exam handwriting
-
Add recogniser v3 (crnn_m_v3_owner): v2.3 fine-tuned on real Khmer exam handwriting
-
Add recogniser v3 (crnn_m_v3_owner): v2.3 fine-tuned on real Khmer exam handwriting
5 days ago
-
Add config.json so Hub download stats are counted
-
Model card: new intro, scores and quick start; add base_model and datasets metadata; library_name onnx
6 days ago
-
Model card: structure v3 + table-structure models, table datasets
-
Add v3 models: structure v3 (tables heads) + table-structure YOLO11n (TSR)
8 days ago
-
Add Models table with links to each version
10 days ago
-
Model card: files live in this repo; add Khmer OCR search tags
-
Restore model files into this repo (single combined ThaOCR repo)
Re-adds layout/, textdet/, rec/, font/, structure/ files exactly as in commit a4711e1.
11 days ago
-
Release v3: index only (model files live in the stage repos), add Versions section
-
Update index: layout v4 default, licence section, manifest with v4
-
layout v4 nano (stamps, logos, handwriting, letterheads) + int8 variant
-
ThaOCR Khmer document OCR models v2 (layout, textdet, rec, font, structure)
-
Turn into index repo: model files moved to one repo per stage
-
ThaOCR Khmer document OCR models v2
-
initial commit