Results and benchmarks
PP-OCRv2: Bag of Tricks for Ultra Lightweight OCR System is the primary contribution described in this paper.
| Task | Dataset | Metric | Value | Source |
|---|---|---|---|---|
| Bag Tricks Ultra Lightweight Ocr System | MobileNetV1-0.75x | Top1-Acc | 68.81 | paper-derived |
| Bag Tricks Ultra Lightweight Ocr System | MobileNetV2-0.75x | Top1-Acc | 69.83 | paper-derived |
| Bag Tricks Ultra Lightweight Ocr System | MobileNetV3-small-1.0x | Top1-Acc | 68.24 | paper-derived |
Audit each benchmark finding before selecting an implementation path. Evidence refs map to the disclosure below.
Evidence graph: 4 refs, 4 links.
Utility signals: depth 90/100, grounding 95/100, status high.
Implementation
Best maintained implementation now
Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.
88,221 stars · 11,223 forks · Last push Jul 22, 2026 · Apache-2.0 license
- License
- CI
- Dependencies
- Docker
Official implementation from Papers with Code · Repository link is mentioned in the paper metadata · Strong overlap with paper title keywords
PaddlePaddle/PaddleOCR is the strongest maintained implementation based on ranking signals. CI workflows are present. License is declared (Apache-2.0).
Open PaddlePaddle/PaddleOCR- No repository-level red flags were detected, but paper-specific preprocessing and hyperparameter details may still be under-specified.
- Selected PaddlePaddle/PaddleOCR as the strongest maintained implementation for new work.
- Includes CI workflow signals.
- Includes dependency/environment manifest signals.
- Repository activity is within the last 24 months.
Compare implementation paths
Compare maintenance quality, reproducibility coverage, and evidence confidence before choosing a reproduction baseline.
- Maintenance
- Recently updated
- Confidence
- High
- Reproducibility
- Strong
- Stars
- 88,221
- Last push
- Jul 22, 2026 (35d)
Official implementation from Papers with Code · Repository link is mentioned in the paper metadata
- No Docker setup
- Maintenance
- Recently updated
- Confidence
- Low
- Reproducibility
- Limited
- Stars
- 1,205
- Last push
- Jul 10, 2026 (46d)
Community adoption signal (1205 stars)
- No CI pipeline detected
- No tagged releases
- No Docker setup
- Maintenance
- Recently updated
- Confidence
- Low
- Reproducibility
- Strong
- Stars
- 0
- Last push
- Jul 24, 2026 (32d)
Matched via arXiv identifier search
- No tagged releases
- Low confidence match
Reproduction readiness
Ready to run
Ready to reproduce
- Clone PaddlePaddle/PaddleOCR and install dependencies from pyproject.toml.
- CI pipeline detected, so automated tests are in place.
- Last updated 35 days ago.
Quick start
git clone https://github.com/PaddlePaddle/PaddleOCR.git
pip install -e .Repositories and ecosystem
No additional verified repositories beyond the primary recommendation.
These repositories had low-confidence matching signals and are hidden by default.
- frotms/PaddleOCR2Pytorch
Confidence: Low · 1,205 stars
- BinayakBM0712/VeriDoc-AI
Confidence: Low · 0 stars
- Parsabzh/Paddle_convert
Confidence: Low · 1 stars
Hugging Face artifacts
No direct paper-linked artifacts were found. Showing strongest curated related artifacts for faster exploration.
Models
- BAAI/bge-reranker-v2.5-gemma2-lightweight
164,577 downloads · 55 likes
Broaden model search
Datasets
No trustworthy datasets matches right now.
Search datasets on Hugging FaceSpaces
No trustworthy spaces matches right now.
Search spaces on Hugging FaceResearch context
Tasks
Bag Tricks Ultra Lightweight Ocr System
Methods
None detected
Domains
None detected
Open this paper in HFEPX to review benchmark signals, evaluation modes, and human-feedback protocol context.
Open in HFEPXJump to Paper2Code search queries derived from this paper's research context.
Data includes links from Papers with Code ( CC-BY-SA-4.0 ).